Paraffin-embedded tissue specimens from 343 patients with confirmed NSCLC (including 125 EGFR Del 19, 100 L858R mutations as well as 118 wild-type EGFR selected by ARMS, and the latter two as comparisons) and 30 corresponding normal lung tissues, collected from 2010 to 2013, were analyzed from an archived thoracic oncology tissue repository at the Department of Thoracic Surgery of Tangdu Hospital affiliated with The Fourth Military Medical University. Patients who received preoperative chemotherapy, radiotherapy or EGFR-targeted therapy were excluded from the present study. Detailed information was collected from the medical records of the enrolled patients in a computerized registry database including patient age, gender, smoking history, clinical manifestation, surgical method, tumor status, histological differentiation, nodal status and follow-up information. The follow-up lasted until September 30, 2014, with a median follow-up period of 21.65 months for living patients (range, 7.43–48.17 months). The day of surgery was considered as the starting day for estimating postoperative survival time. Histological classification of tumors was reviewed by pathologists and based on the World Health Organization Criteria. All tumors were staged according to the pathological tumor/node/metastasis (pTNM) classification (7th edition) of The World Health Organization. The study protocol was approved by the Regional Ethics Committee for Clinical Research of the Fourth Military Medical University. All patients provided written informed consent for use of their medical records and tissue specimens for research purposes.

The tumor samples were fixed with 10% formaldehyde and embedded with paraffin. Sections were sliced to 4-µm thickness, deparaffinized with a series of xylene and rehydrated through a graded series of alcohol. Microwave antigen retrieval was performed at 750 W for 5 min and 450 W for 15 min in citrate buffer (pH 6.0) to enhance the immunoreactivity. After blocking the endogenous peroxidase activity with 3% hydrogen peroxidase for 30 min, the sections were incubated with 5% normal goat serum for 30 min at room temperature to block non-specific antibody reaction. After washing the tissue samples with phosphate-buffered saline (PBS) three times for 5 min, the sections were incubated with the primary antibodies [EGFR Del 19 at 1:100 (Cell Signaling Technology, Beverly, CA, USA); Fn14 at 1:100 (Abcam, Cambridge, UK); p-JAK1 at 1:100; p-STAT1 at 1:100 (both from Abcam)] overnight at 4°C, and incubated with an EnVision™ Detection kit (Dako, Glostrup, Denmark) following the manufacturer's instructions. The sections were then reacted with 0.003% 3,3′-diaminobenzidine and counterstained with hematoxylin. To confirm the specificity of the immunostaining, negative controls were obtained by replacing the primary antibody with PBS.

Five random fields from each section were viewed under a light microscope (Leica DM4000B; Leica, Wetzlar, Germany) in a high-power field (magnification, x400). We scored all sections in accordance with previous studies on IHC staining (18,19). Briefly, a total immunostaining score was calculated as the product of a proportion score and an intensity score. The proportion score represented the estimated fraction of positively stained tumor cells (0, 0–5; 1, 6–25; 2, 26–50; 3, 51–75 and 4, 76–100%). The intensity score represented the estimated staining intensity (0, negative; 1, weak; 2, moderate; 3, strong). The score of these were measured according to the result of the degree multiplied by the score of the staining intensity: 0, 0; 1+, 1–4; 2+, 5–8 and 3+, 9–12. Score 0 was considered negative, whereas scores (1+ to 3+) were considered positive. Thus, the total score ranged from 0 to 12. All slides were assessed by two independent investigators who were blinded to the clinical features and outcomes. The final IHC staining score reported is the average of the scores from the two investigators.

RT-qPCR was performed for gene expression detection in specimens of NSCLC with EGFR Del 19 and corresponding normal lung tissues. Total RNA was isolated using the E.Z.N.A. Total RNA Kit I Bio-Tek (R6834; omega Bio-Tek, Norcross, GA, USA) and reverse-transcribed to cDNA using RevertAid First Strand cDNA Synthesis kit (K1622; Thermo Fisher Scientific, Vilnius, Lithuania). The primers were as follows: forward primer, 5′-CCAAGCTCCTCCAACCACAA-3′ and reverse primer, 5′-TGGGGCCTAGTGTCAAGTCT-3′ for Fn14 gene; forward primer, 5′-GGGAAATCTGCTACAATGGC-3′ and reverse primer, 5′-TGATGGCTCGGAAGAAAGGC-3′ for JAK1 gene; forward primer, 5′-GTTGAACCCTACACGAA-3′ and reverse primer, 5′-TACAGAGCCCACTATCC-3′ for STAT1 gene; forward primer, 5′-GAGCTACGAGCTGCCTGACG-3′ and reverse primer, 5′-CCTAGAAGCATTTGCGGTGG-3′ for β-actin gene. β-actin served as the internal control.

Cytoplasmic and nuclear protein were extracted from the tissues using RIPA lysate (P0013B; Beyotime Institute of Biotechnology, Jiangsu, China) and western blotting was performed using the EGFR Del 19 monoclonal antibody (6B6; Cell Signaling Technology), Fn14-specific monoclonal antibody (EPR3179), p-JAK1 monoclonal antibody (EPR1899), p-STAT1 monoclonal antibody (EPR3146) (all from Abcam) and the β-actin-specific polyclonal antibody (CW0097; CW Biotech Co., Ltd., Beijing, China).

Protein concentrations were determined using the BCA assay kit (Pierce Biotechnology, Inc., Rockford, IL, USA). Equivalent amounts of each protein sample were mixed with loading buffer (CW0027A; CWBio, Beijing, China), heated at 65°C for 30 min, and subjected to SDS-PAGE using 12% separation and 5% spacer gels. Then the protein was transferred to a polyvinylidene fluoride (PVDF) membrane (Solarbio, Beijing, China) by electroblotting (Bio-Rad, Philadelphia, PA, USA). The membrane was blocked for 3 h at room temperature in Tris-buffered saline and Tween-20 (TBST) (25 mM Tris/HCl, pH 7.5, 150 mM NaCl, 0.1% Tween-20) containing 5% non-fat dry milk. The membranes were incubated overnight at 4°C in 1:2,000 dilution of anti-EGFR Del 19 monoclonal antibody, 1:5,000 dilution of anti-Fn14 monoclonal antibody, 1:5,000 dilution of p-JAK1 monoclonal antibody, 1:7,500 dilution of p-STAT1 monoclonal antibody and 1:2,500 dilution of β-actin polyclonal antibody with WB Antibody Diluent (P0023A; Beyotime Institute of Biotechnology), washed six times in TBST, and incubated for 35 min in a 1:5,000 dilution of goat anti-rabbit Ig-HRP (EK020; Zhuangzhi Bio, Xi'an, China) with WB Secondary Antibody Diluent (P0023D; Beyotime Institute of Biotechnology). Immunoreactive bands were revealed by the enhanced chemiluminescence system (Santa Cruz Biotechnology, Santa Cruz, CA, USA). Images were captured and analyzed by GelDox XR system (Bio-Rad).

The tumor samples were fixed with 10% formaldehyde and embedded with paraffin. Sections were sliced to 4-µm thickness, deparaffinized with a series of xylene and rehydrated through a graded series of alcohol. Microwave antigen retrieval was performed at 750 W for 5 min and 450 W for 15 min in citrate buffer (pH 6.0) to enhance the immunoreactivity. After blocking the endogenous peroxidase activity with 3% hydrogen peroxidase for 30 min, the sections were incubated with 5% normal goat serum for 30 min at room temperature to block non-specific antibody reaction. After washing the tissue samples with PBS three times for 5 min, the sections were incubated with the primary antibodies [EGFR Del 19 at 1:100 (Cell Signaling Technology); Fn14 at 1:100; p-JAK1 at 1:100; p-STAT1 at 1:100 (all from Abcam)] overnight at 4°C. After an additional series of washes, the slides were stained with goat anti-rabbit (Cy3; Zhuangzhi Bio, Xi'an, China) at room temperature for 50 min, and then the tissue samples were washed with PBS three times for 5 min. The sections were then incubated with the other primary antibodies overnight at 4°C. After an additional series of washes, the slides were stained with goat anti-mouse (Alexa Fluor 488; Zhuangzhi Bio) at room temperature for 50 min. After the final washing, the slides were mounted in 50% glycerol (in PBS) and examined by a fluorescence microscope (Leica DM4000B).

As our data were not normally distributed, we used non-parametric tests. The degree of protein expression in the primary tumor was compared by cross-table analysis. The expression differences among more than three groups were analyzed by Kruskal-Wallis H test, otherwise Mann-Whitney U analysis was used. Survival curve was examined using the Kaplan-Meier method, and correlation analyses of the survival time and various clinicopathological variables were performed by univariate and multivariate analyses using the Cox regression model. All analyses were performed using SPSS 15.0 software (SPSS, Inc., Chicago, IL, USA). P-values were adjusted for multiple testing and a P<0.05 was considered to indicate a statistically significant result. Spearman's rank correlation coefficients were calculated for the assessment of overall concordance, and r_s was used to express the relevant coefficient, and when P<0.05, r_s>0.3 was considered positive relevance.

The clinicopathological characteristics of the NSCLC patients with EGFR Del 19 are summarized in Table I. There were 62 female and 63 male patients with a median age of 58 years (range, 36–79 years). The patients were diagnosed with squamous cell carcinoma (SCC) (n=30, 24%) and adenocarcinoma (ADC) (n=95, 76%). Histopathologic diagnosis included: well differentiated (n=23, 18.4%), moderately differentiated (n=83, 66.4%) and poorly differentiated (n=19, 15.2%) tumors. Postoperative staging evaluation demonstrated stage I disease in 43 patients, stage II disease in 19 patients, stage III disease in 57 patients and stage IV disease in 6 patients.

In the present study, 99.2% (124/125) of the tumor sections were classified as EGFR Del 19-positive as detected by IHC and positive staining was mainly located in the cytoplasm (Fig. 1A and B). The correlation of EGFR Del 19 expression with clinicopathological parameters was then investigated. EGFR Del 19 expression was significantly associated with differentiation (P<0.001), pTNM stage (P=0.001), primary tumor size (P=0.008), lymph node metastasis (P=0.003) and tumor location (P=0.032). No significant relationship was noted between EGFR Del 19 expression and gender (P=0.563), age (P=0.911), smoking history (P=0.626) and histological type (P=0.480) (Table I).

Differences in the expression of Fn14 were detected in NSCLC with EGFR Del 19 (n=125) and its corresponding normal lung tissues (n=30), EGFR L858R samples (n=100) and EGFR wild-type NSCLC samples (n=118). A total of 100% (125/125) of tumor sections were classified as Fn14-positive while 100% (30/30) corresponding normal lung tissue sections were classified as Fn14-negative. Seventy-two percent (72/100) and 48.3% (57/118) of the tumor sections were classified as Fn14-positive in L858R EGFR and wild-type EGFR NSCLC samples, nevertheless the rate of high expression of Fn14 (++−+++) in Del 19 EGFR, L858R EGFR and wild-type EGFR NSCLC samples was 65.6, 17 and 12.7%, and there was significant difference in Fn14 expression among EGFR Del 19, L858R EGFR, and wild-type EGFR samples (P<0.001). Thus, Fn14 was highly expressed in NSCLC with EGFR Del 19 (Table II).

Fn14-positive staining was mainly located on the cell membrane and in the cytoplasm (Fig. 1C and D). In order to evaluate the role of Fn14 in NSCLC, the correlation of Fn14 expression with clinicopathological parameters was investigated (Table I). The results showed that Fn14 expression was significantly associated with differentiation (P<0.001), pTNM stage (P<0.001), primary tumor size (P=0.029), lymph node metastasis (P<0.001) and tumor location (P= 0.031). No significant relationship was noted between Fn14 expression and gender (P=0.282), age (P=0.498), smoking history (P=0.217) and histological type (P=0.080).

p-JAK1 (83.2%, 104/125) (Fig. 1E and F) and p-STAT1 (100%, 125/125) (Fig. 1G and H) were mainly located in the cytoplasm and cell nucleus. Correlation of JAK1 and STAT1 expression with clinicopathological parameters was then investigated. p-JAK1 and p-STAT1 expression was significantly associated with differentiation (P<0.01), pTNM stage (P<0.001) and lymph node metastasis (P<0.001). No significant relationship was noted with gender, age, smoking history, histological type, primary tumor size and tumor location (Table I).

The mRNA expression levels of Fn14, JAK1 and STAT1 were detected by RT-qPCR. Compared to corresponding normal lung tissues, they were increased in the NSCLC with EGFR Del 19 tissues (Fig. 2). In addition, the protein expression levels of EGFR Del 19, Fn14, p-JAK1 and p-STAT1 were detected by western blotting in NSCLC with EGFR Del 19 and the corresponding normal lung tissues. The results showed that when compared to the normal lung tissues, Fn14 and p-STAT1 were increased in NSCLC, but this trend was less evident for p-JAK1 protein expression (Fig. 3).

In order to investigate the correlation of EGFR Del 19 with Fn14, p-JAK1, p-STAT1 expression, the expression rate and intensity of these proteins were analyzed (Table III). The correlation of EGFR Del 19 with Fn14, p-JAK1 and p-STAT1 expression was obvious (P<0.05; r_s>0.3) in NSCLC, ADC and SCC. The correlation of EGFR Del 19 with Fn14, p-JAK1 and p-STAT1 was high in moderate pathologic stage tissues (P<0.01; r_s>0.423). In addition, there were good correlations for TNM stage in NSCLC (P<0.05; r_s>0.376).

In order to understand the role of Fn14 in NSCLC with EGFR Del 19 and the influence of Fn14 on downstream signaling pathway, correlations between Fn14 expression with JAK1/STAT1 signaling molecules were analyzed (Table IV). There were obvious correlations in ADC and SCC (P<0.05; r_s>0.3). Correlations between Fn14 and JAK1 and STAT1 were strong in moderately differentiated tissues (P<0.01; r_s>0.366). In addition, there was a good correlation with pTNM stage in NSCLC (P<0.01; r_s>0.336).

EGFR Del 19 and Fn14 were mainly expressed on the cell membrane. In order to further verify whether there is a co-localization phenomenon when they are expressed, immunofluorescence experiments were conducted. Preliminary experimental results showed that the expression of localization phenomenon existed. p-JAK1 and p-STAT1 were mainly located in the cytoplasm and cell nucleus. In addition, the same method was used for co-localization of p-JAK1 and p-STAT1 protein expression. Co-localization phenomenon also existed (Fig. 4).

Among factors such as patient age, gender, smoking history, tumor histological type, lymph node metastasis, differentiation and pTNM stages, only the pTNM stage, differentiation and lymph node metastasis were significantly associated with patient survival. The median survival time of patients with pTNM I/II tumors (n=61) was not reached [95% confidence interval (CI) was not reached], whereas the median survival time of those with TNM III/IV tumors (n=64) was 24 months (95% CI, 17.96–30.04 months; P<0.001; Fig. 5A). Patients with well or moderately differentiated tumors (n=140, median survival time, 37.47 months, 95% CI, 31.73–43.20 months) had a longer survival time than those with poorly differentiated tumors (n=51, median survival time, 13 months; 95% CI, 8.67–17.34 months; P<0.001; Fig. 5B). The median survival time of patients with no lymph node metastasis (n=74) was not reached (95% CI was not reached), whereas the median survival time of those with lymph node metastasis (n=51) was 32.07 months (95% CI of ratio, 1.382–4.733; P<0.01; Fig. 5C).

To investigate the relationship between EGFR Del 19 and Fn14 expression and the clinical outcome of NSCLC patients, the correlation of EGFR Del 19 and Fn14 expression with patient survival and EGFR Del 19, Fn14 expression status were analyzed. Patients with Fn14 high positive expression had a significantly worse prognosis than those with Fn14 low positive expression (P=0.001; Fig. 5D). The median survival time of patients with Fn14 high positive NSCLC (n=85) was 28 months (95% CI, 17.99–38.01 months) and the mean survival time was 27.07 months (95% CI, 23.49–30.66 months), whereas the median survival time of patients with Fn14 low positive NSCLC (n=40) was not reached and the mean survival time was 33.86 months (95% CI, 31.17–36.55 months). Patients with high positive expression of EGFR Del 19 had a significantly worse prognosis than those with low positive expression of EGFR Del 19 (P=0.0155; Fig. 5E). The median survival time of patients with EGFR Del 19 low positive NSCLC (n=52) was 32.8 months (95% CI of ratio, 0.2551–0.8663), whereas the median survival time of patients with EGFR Del 19 high positive NSCLC (n=73) was not reached. Comprehensive statistics were analyzed concerning the influence of EGFR Del 19 and Fn14 on patient prognosis. We found that with increased expression of EGFR Del 19 and Fn14, the shorter the patient survival (P<0.01; Fig. 5F).

To further assess whether EGFR Del 19 and Fn14 expression represents a common prognostic parameter in NSCLC patients with EGFR Del 19, regression analysis using the Cox's proportional hazards model was performed. The covariate parameters included several clinicopathological factors in addition to EGFR Del 19 and Fn14, as shown in Table V. In the univariate analysis, factors including EGFR Del 19 expression, Fn14 expression, pTNM stage and differentiation showed a significantly higher hazard ratio for poor prognosis.

Moreover, multivariate analysis was carried out using the significant factors observed in the univariate analysis. The results showed that, in addition to TNM stage and differentiation, Fn14 expression was an independent prognostic factor (P=0.047; Table V). These results strongly indicate that Fn14 expression in NSCLC patients is closely related to poor prognosis.