General conditions
There were 149 patients with NSCLC, 151 patients with tuberculosis and 150 healthy controls in this research. Characteristics of patients grouped were showed in Table 1. There were no statistical differences in gender, age and comorbidities among these three groups (P = 0.929 and P = 0.573, respectively, Table 1). Nevertheless, NSCLC patients showed higher smoking rate (current: 43.62%, former: 36.91%, never: 19.46%) and tobacco exposure (42.65 ± 16.94 years) compared with healthy controls (current: 30%, former: 26.67%, never: 43.33%, P = 0.000; 30.23 ± 9.12 years, P = 0.000, respectively) and tuberculosis patients (current: 31.13%, former: 39.07%, never: 29.8%, P = 0.039; 36.78 ± 12.97 years, P = 0.000, respectively, Table 1)
Table 1
The comparison of clinical characteristics among NSCLC, tuberculosis and healthy control.
|
Healthy control (n = 150)
|
Tuberculosis
(n = 151)
|
NSCLC
(n = 149)
|
P value
|
Age (years)
|
59.92 ± 8.33
|
59.27 ± 9.70
|
60.08 ± 9.94
|
0.734
|
Gender
|
|
|
|
0.932
|
male
|
116 (77.33%)
|
114(75.5%)
|
114(76.51%)
|
|
female
|
34(22.67%)
|
37(24.5%)
|
35(23.49%)
|
|
Smoking status
|
|
|
|
< 0.001
|
Current
|
45(30%)
|
47(31.13%)
|
65(43.62%)
|
|
Former
|
40(26.67%)
|
59(39.07%)
|
55(36.91%)
|
|
Never
|
65(43.33%)
|
45(29.8%)
|
29(19.46%)
|
|
Tobacco Exposure (years)
|
30.23 ± 9.12
|
36.78 ± 12.97
|
42.65 ± 16.94
|
< 0.001
|
Comorbidities
|
|
|
|
|
Hypertension
|
55(36.67%)
|
54(35.76%)
|
60(40.27%)
|
0.696
|
Diabetes
|
15(10%)
|
12(7.95%)
|
18(12.08%)
|
0.491
|
Asthma
|
7
|
9(5.96%)
|
12(4.03%)
|
0.473
|
The cfDNA and SUV-Maxa were increased in NSCLC patients
The serum level of cfDNA [19.78(11.52, 28.36) ng/µl] and SUV-Maxa [3.99 (2.33, 5.71)] were significantly higher in patients with NSCLC than those in healthy controls [9.75 (5.27, 13.65) ng/µl, P < 0.001, Fig. 1B; 1.93 (1.05, 2.59), P < 0.001, Fig. 1C, respectively] and those in patients with tuberculosis [14.58 (9.149, 18.74) ng/µl, P < 0.001, Fig. 1B; 2.41 (1.50, 3.12), P < 0.001, Fig. 1C, respectively].
The cfDNA levels were associated with SUV-Max a in NSCLC patients
In healthy controls and patients with tuberculosis, no statistical association was found between SUV-Maxa and age (P = 0.319 and P = 0.102, respectively), in contrast, statistical association was found in NSCLC patients (P = 0.029, Table 2). In addition, adenocarcinoma of NSCLC patients showed higher cfDNA and SUV-Maxa than those in squamouscarcinoma [20.93 (16.89, 36.23) ng/µl vs 17.37 (8.82, 22.81) ng/µl, P = 0.000; 4.36 (3.47, 8.05) vs 3.01 (1.57, 4.22), P = 0.000, respectively]. Moreover, patients with advanced NSCLC (pathology stage Ⅲ-Ⅳ) owned higher cfDNA and SUV-Maxa than those in early NSCLC patients (pathology stage Ⅰ-Ⅱ) [21.57 (15.95, 31.05) ng/µl vs 17.39 (7.95, 20.56) ng/µl, P = 0.001; 4.31 (3.72, 7.95) vs 2.89 (1.35, 3.46), P = 0.000, respectively, Table 2]. Similarly, there is no significant correlation of cfDNA and SUV-Maxa in healthy controls (r = 0.163, P = 0.045, Fig. 1D) and patients with tuberculosis (r = 0.226, P = 0.005, Fig. 1E). However, patients with NSCLC showed obvious correlation of cfDNA and SUV-Maxa (r = 0.841, P < 0.001, Fig. 1F).
Table 2
Correlation between clinical characteristics and serum cfDNA or SUVMaxa
|
N
|
cfDNA
|
K-Wχ 2 /Z
|
P
|
SUVMaxa
|
K-Wχ 2 /Z
|
P
|
Healthy control
|
|
|
|
|
|
|
|
Age (years)
|
|
|
|
|
|
|
|
≤ 65
|
119
|
9.58 (5.06, 12.54)
|
0.991
|
0.319
|
1.88 (1.14, 2.55)
|
0.501
|
0.479
|
> 65
|
31
|
10.01 (7.15, 17.75)
|
|
|
1.94 (0.89, 2.71)
|
|
|
Gender (N)
|
|
|
|
|
|
|
|
Male
|
116
|
9.82 (6.48, 12.53)
|
-0.608
|
0.543
|
1.90 (1.09, 2.41)
|
0.514
|
0.607
|
Female
|
34
|
7.14 (3.52, 17.62)
|
|
|
2.14 (.79, 3.41)
|
|
|
Tuberculosis
|
|
|
|
|
|
|
|
Age (years)
|
|
|
-0.102
|
0.918
|
|
-1.448
|
0.148
|
≤ 65
|
117
|
14.59 (11.34, 17.91)
|
|
|
2.53 (1.69, 3.12)
|
|
|
> 65
|
34
|
14.62 (6.30, 26.99)
|
|
|
2.10 (1.42, 3.25)
|
|
|
Gender (N)
|
|
|
-0.311
|
0.755
|
|
-0.270
|
0.787
|
Male
|
114
|
14.50 (9.14, 18.78)
|
|
|
2.43 (1.52, 2.99)
|
|
|
Female
|
37
|
14.66 (9.73, 18.94)
|
|
|
2.23 (1.46, 5.15)
|
|
|
NSCLC
|
|
|
|
|
|
|
|
Age (years)
|
|
|
-1.636
|
0.102
|
|
-2.182
|
0.029
|
≤ 65
|
113
|
18.64 (14.45, 23.64)
|
|
|
3.53 (2.86, 4.72)
|
|
|
> 65
|
36
|
27.90 (6.68, 41.91)
|
|
|
7.19 (2.04, 8.94)
|
|
|
Gender (N)
|
|
|
-1.110
|
0.267
|
|
-1.157
|
0.247
|
Male
|
114
|
19.05 (11.17, 26.12)
|
|
|
3.81 (2.28, 5.33)
|
|
|
Female
|
35
|
20.29 (14.29, 35.03)
|
|
|
4.05 (3.00, 8.30)
|
|
|
Pathological type
|
|
|
-3.682
|
< 0.001
|
|
-4.080
|
< 0.001
|
Adenocarcinoma
|
76
|
20.93 (16.89, 36.23)
|
|
|
4.36 (3.47, 8.05)
|
|
|
Squamous Carcinoma
|
73
|
17.37 (8.82, 22.81)
|
|
|
3.01 (1.57, 4.22)
|
|
|
pathology stage
|
|
|
-3.258
|
0.001
|
|
-5.660
|
< 0.001
|
StageⅠ-Ⅱ
|
64
|
17.39 (7.95, 20.56)
|
|
|
2.89 (1.35, 3.46)
|
|
|
Stage Ⅲ-Ⅳ
|
85
|
21.57 (15.95, 31.05)
|
|
|
4.31 (3.72, 7.95)
|
|
|
NSCLC: non-small cell lung cancer; cfDNA : cell-free DNA; SUV-Maxa: the maximum standardized uptake value
|
Combining cfDNA and SUV-Max a to distinguish NSCLC from healthy controls
ROC curve analysis showed that the combination of cfDNA and SUV-Maxa (AUC = 0.982, P = 0.000, cut-off values = 0.22, sensitivity = 94.5%, specificity = 92.7%) displayed higher efficacy to distinguish NSCLC from healthy controls than alone use (AUC = 0. 907, P = 0.000, cut-off values = 20.25, sensitivity = 98.0%, specificity = 67.3%; AUC = 0.901, P = 0.000, cut-off values = 7.67, sensitivity = 80.6%, specificity = 94.3%, respectively, Fig. 1G, Table 3).
Table 3
Combining cfDNA and SUVMaxa to distinguish NSCLC from healthy controls.
|
AUC ROC
|
Cut off value
|
Sensitivity (%)
|
Specificity (%)
|
95% CI
|
P value
|
cfDNA
|
0.907
|
20.25
|
98.0
|
67.3
|
0.875–0.939
|
< 0.001
|
SUVMaxa
|
0.901
|
7.67
|
80.6
|
94.3
|
0.860–0.941
|
< 0.001
|
combination of cfDNA and AUCMaxa
|
0.982
|
0.22
|
94.5
|
92.7
|
0.971–0.992
|
< 0.001
|
NSCLC: non-small cell lung cancer; cfDNA : cell-free DNA; SUVMaxa: the maximum standardized uptake value; AUC: area under the curve
|
Combining cfDNA and SUV-Maxa to distinguish NSCLC from tuberculosis
ROC curve analysis showed that the combination of cfDNA and SUV-Maxa (AUC = 0.935, P = 0.000, cut-off values = 0.46, sensitivity = 91.9%, specificity = 90.1%) also displayed better effect to distinguish NSCLC from tuberculosis than alone use (AUC = 0.804, P = 0.000, cut-off values = 16.83, sensitivity = 96%, specificity = 65.3%; AUC = 0.851, P = 0.000, cut-off values = 3.51, sensitivity = 67.1%, specificity = 93.7%; respectively, Fig. 1H, Table 4).
Table 4
Combining cfDNA and SUVMaxa to distinguish NSCLC from tuberculosis.
|
AUC ROC
|
Cut off value
|
Sensitivity (%)
|
Specificity (%)
|
95% CI
|
P value
|
cfDNA
|
0.804
|
16.83
|
96
|
65.3
|
0.751–0.857
|
< 0.001
|
SUVMaxa
|
0.851
|
3.51
|
67.1
|
93.7
|
0.807–0.896
|
< 0.001
|
combination of cfDNA and AUCMaxa
|
0.935
|
0.46
|
91.9
|
90.1
|
0.910–0.960
|
< 0.001
|
NSCLC: non-small cell lung cancer; cfDNA : cell-free DNA; SUVMaxa: the maximum standardized uptake value; AUC: area under the curve
|
The cfDNA was associated with [3H]-2-deoxy-DG uptake in NSCLC mice model
To investigate the specific mechanism of the relationship between cfDNA and SUV-Maxa, Two NSCLC mice models (A549 and NCI–H460) and the tuberculosis mice model were used. Metabolic tumor burden was displayed by the value of SUV-Maxa in PET/CT, which depended on the rate of contrast agent (18F-FDG) uptake by tumor [16, 17]. Therefore, the rate of [3H]-2-deoxy-DG uptake was measured to show the metabolic tumor burden in lung tumor of NSCLC mice model. We found A549 NSCLC group and NCI–H460 NSCLC group both had a higher rate of [3H]-2-deoxy-DG uptake than those in control group and in tuberculosis group (Fig. 2B), similarly to human. Moreover, NSCLC group also showed higher cfDNA level than those in the control group and tuberculosis group (Fig. 2A). Similarly, there is also no correlation of cfDNA and the rate of [3H]-2-deoxy-DG uptake in the control group (Fig. 2C) and tuberculosis group (Fig. 2D). However, NSCLC group showed obvious association of cfDNA level and [3H]-2-deoxy-DG uptake (Fig. 2E, Fig. 2F).
GLU1/FasL/Caspase-8/Caspase-3 were upregulated in lung tumor tissue of NSCLC mice model
To further research the specific mechanism of glucose uptake of lung tumor interacting with serum cfDNA, RNA microarray analysis was performed. GLU1, FasL, Caspase-8, caspase-3 gene were significantly upregulated in lung tumor of A549/NCI–H460 NSCLC group than those in tuberculosis group (Fig. 2G). Besides that, Western blotting also showed that GLU1, FasL, cleaved-caspase 8 and cleaved-caspase 3 were significantly increased, while pro-caspase 8 and pro-caspase 3 were decreased in A549/NCI–H460 NSCLC group than those in tuberculosis group (Fig. 2H-I).
GLU1 promoted the [3H]-2-deoxy-DG uptake in NSCLC
To investigate whether [3H]-2-deoxy-DG uptaking rely on GLU1, the GLU 1 inhibitor (WZB117) was used. The increased [3H]-2-deoxy-DG uptake was reversed by WZB117, indicating that the upregulated GLU1 promoted the [3H]-2-deoxy-DG uptake in A549/NCI–H460 NSCLC (Fig. 3C, G).
GLU1 increased cfDNA levels by FasL/caspase 8/caspase 3 pathway in vivo
To explore whether the necroptosis-induce cfDNA was mediated through GLU1/FasL/caspase 8/caspase 3 pathways, GLU 1 inhibitor (WZB117), anti–FasL antibody, caspase 8 and caspase 3 inhibitor (Z-IETD-FMK) were used in vivo. We found that GLU1 increased the expression of cleaved-caspase 3 through increasing FasL and cleaved-caspase 8 expression in lung tumor of A549/NCI–H460 NSCLC mice (Fig. 3A,D,E,H), indicating that GLU1/FasL/caspase 8/caspase 3 pathway was activated in NSCLC to promote necroptosis. Moreover, the increased serum cfDNA was reversed by WZB117, anti–FasL antibody and Z-IETD-FMK (Fig. 3B,F), indicating that GLU1 increased cfDNA levels by FasL/caspase 8/caspase 3 pathways in vivo.
GLU1 promoted apoptosis by FasL/caspase 8/caspase 3 pathway in vitro
Vitro experiments also demonstrated the similar results in A549 and NCI–H460 cells line. Knocking down GLU1 and FasL gene by siRNA and inhibited caspase 8/caspase 3 activities could significantly reduce apoptosis, indicated that GLU1 promote apoptosis by FasL/caspase 8/caspase 3 pathway in vitro (Fig. 4).