Expression of CENPM in lung adenocarcinoma
Using the TCGA database, we analyzed the CENPM gene expression in various tumors and normal tissue samples. We found that CENPM gene was highly expressed in almost all tumor types, including lung adenocarcinoma (n=515) (Figure 1A). CENPM was found to be highly expressed in lung adenocarcinoma by analyzing all the differentially expressed genes between lung adenocarcinoma and normal tissues from the database (Figure 1B). From the TCGA database comprised of 535 lung adenocarcinoma samples and 59 normal tissue samples, we could see that CENPM expression was significantly higher in lung adenocarcinoma tissues (3.628 VS. 2.227, P < 0.001; Figure 1C, Table 1). Consistently, CENPM was highly expressed in lung adenocarcinoma samples from the TCGA database compared with normal tissue samples from the GTEx database (3.628 vs. 2.139, p < 0.001; Figure 1D). Next, we analyzed 57 matched lung adenocarcinomas and normal tissues from the TCGA database and found that the CENPM expression in tumor tissues was significantly higher than that in paired normal tissues (3.709 VS. 2.234, p < 0.001; Figure 1E, Table 1).
These findings were further validated using combined data from three independent datasets derived from GSE43458, GSE32863, and GSE140797 datasets. CENPM expression was significantly higher in lung adenocarcinoma (n=145) as compared with normal tissue samples (n=95) (3.012 vs. 2.901, p<0.001; Figure 1F, Table S1). Next, we compared 65 paired tumor and normal samples from the same dataset and saw consistent results that lung adenocarcinoma tissues had higher CENPM expression than their paired adjacent normal lung tissues (3.089 vs 2.921, p<0.001; Figure 1G, Table S1).
Next, we analyzed the diagnostic value of CENPM expression for lung adenocarcinoma. The area under the ROC curve was 0.874 (95% CI: 0.834 – 0.913). From the results, it can be seen that the predictive ability of CENPM expression was accurate in distinguishing between tumor and normal tissues. Using an optimal CENPM gene expression cut-off of 2.919, the sensitivity was 86.4% and specificity was 78.9% in distinguishing between tumor and normal tissues (Figure 1H). Taken together, our analysis showed that CENPM was highly expressed in lung adenocarcinoma than normal tissues.
Table 1
CENPM expression in TCGA database
| Unpaired Samples | | Paired samples | |
| Normal(n=59) | Tumor(n=535) | p | Normal(n=57) | Tumor(n=57) | p |
Mean | 2.227 | 3.682 | <0.001 | 2.234 | 3.709 | <0.001 |
SD | 0.749 | 0.999 | | 0.761 | 0.968 | |
Max | 4.005 | 6.722 | | 4.005 | 6.16 | |
Min | 0.663 | 0.721 | | 0.663 | 1.901 | |
Cenpm Expression And Clinical Characteristics
Table 2 shows the baseline clinical characteristics of patients with lung adenocarcinoma from the TCGA database. According to 50% of the population, 267 patients had low CENPM expression and 268 patients had high CENPM expression. No statistical difference was found in age, gender, and smoking history between the two groups. However, TNM and clinical stage were significantly different between the two groups.
We also analyzed the mRNA expression of bronchial epithelial cells from smokers with and without lung cancer using the GSE4115 dataset (10). We found that CENPM expression was significantly higher in bronchial epithelial cells from patients diagnosed with lung tumors than patients without lung cancer (2.99 VS. 2.52, p < 0.001; (Figure S1), suggesting that CENPM gene expression is cancer-specific and not due to smoking.
Next, we analyzed the CENPM expression with the baseline clinicopathological status of the patients. The gene expression of CENPM in pathological stage II-IV was significantly higher than that in stage I (3.919 vs. 3.622, p = 0.015, Figure 2A). Similarly, patients with T2-4 had significantly higher CENPM expression than patients with T1 (3.803 vs. 3.444, p < 0.001, Figure 2B), and patients with N1-3 had significantly higher CENPM expression than patients with N0 (3.992 vs. 3.573, p < 0.001, Figure 2C). However, CENPM expression was not statistically different in patients with or without metastasis (4.098 versus 3.716, p = 0.060, Figure 2D). Taken together, these data demonstrate that CENPM expression is positively correlated with advanced stage and is cancer-specific.
Table 2
Characteristic of Baseline
Characteristic | CENPM Low (n=267) | CENPM High(n=268) | p |
Age, median (range) | 67 (60, 73) | 65 (57, 71) | 0.057 |
Gender, n (%) | | | 0.068 |
Female | 157 (58.8%) | 129 (48.1%) | |
Male | 110 (41.2%) | 139 (51.9%) | |
Smoker, n (%) | | | 1.000 |
No | 38 (14.2%) | 37 (13.8%) | |
Yes | 224 (85.8%) | 222 (86.2%) | |
T stage, n (%) | | | 0.002 |
T1 | 108 (20.3%) | 67 (12.6%) | |
T2 | 126 (23.7%) | 163 (30.6%) | |
T3 | 24 (4.5%) | 25 (4.7%) | |
T4 | 8 (1.5%) | 11 (2.1%) | |
N stage, n (%) | | | < 0.001 |
N0 | 191 (36.8%) | 157 (30.3%) | |
N1 | 33 (6.4%) | 62 (11.9%) | |
N2 | 30 (5.8%) | 44 (8.5%) | |
N3 | 0 (0%) | 2 (0.4%) | |
M stage, n (%) | | | 0.022 |
M0 | 180 (46.6%) | 181 (46.9%) | |
M1 | 6 (1.6%) | 19 (4.9%) | |
Pathologic stage, n (%) | | | < 0.001 |
Stage I | 171 (32.4%) | 123 (23.3%) | |
Stage II | 50 (9.5%) | 73 (13.9%) | |
Stage III | 34 (6.5%) | 50 (9.5%) | |
Stage IV | 7 (1.3%) | 19 (3.6%) | |
Cenpm Expression And Prognosis Of Lung Adenocarcinoma
According to CENPM expression from the TCGA database, we analyzed the patient survival time between the two groups. Patients with low CENPM expression achieved significantly longer overall survival (OS) time than those with high CENPM expression (57.5 months vs. 47.5 months, p = 0.001, Figure 3A). The median progression free survival (PFS) was 45.9 months (95%CI: 33.6-83.9) for patients with low CENPM expression and 25.7 months (21.2-36.0) for patients with high expression, which showed that the survival of those with low CENPM expression group would be longer (p<0.001, Figure 3B). The survival analysis was further validated using an independent dataset derived from GSE68465 (11). The data were divided into 2 groups according to CENPM expression, with 148 patients in each group. The OS of CENPM low expression group was 78 months (95%CI: 65-105.4), while that of high expression group was 48 months (95%CI: 40-130). Patients with low CENPM expression achieved a better survival time than those with high CENPM expression (p= 0.033, Figure 3C).
Table 3 shows the treatment response of the two groups. In CENPM low expression group, 183 patients achieved complete response (CR), 4 patients had partial response (PR), 19 patients had stable disease (SD), and 19 patients had progressive disease (PD). The objective response rate (ORR) was 83.2% and disease control rate (DCR) was 91.6% for the low CENPM expression group. In CENPM high expression group, 149 patients were CR, 2 patients were PR, 18 patients were SD, and 52 patients were PD. The ORR was 68.3% and DCR was 76.5% for the high CENPM expression group. Patients with low CENPM expression achieved significantly better ORR and DCR than those with high CENPM expression (p<0.001 and p< 0.001).
Table 4 summarizes the Cox regression analyses of the basic clinical characteristics of patients from the TCGA database for OS. We found that the patient's age, gender, and smoking history did not directly affect the patient's OS in this cohort. T, M, N stage as well as pathological stage were identified as prognostic factors for OS, wherein the more advanced the stage, the worse the OS. CENPM expression was identified as an independent prognostic factor for lung adenocarcinoma (HR=1.665, 95% CI: 1.242-2.233, p< 0.001, Figure 3D). The statistically significant factors based on the Cox regression analyses, including T, M, N stage, and the expression of CENPM, were selected as variables to construct an OS prediction model (Figure 3E). Taken together, our data indicate that high CENPM expression is associated with poor prognosis in lung adenocarcinoma.
Table 3
| Low expression(n=225) | High expression(n=221) | p |
CR | 183(81.4%) | 149(67.4%) | |
PR | 4(1.8%) | 2(0.9%) | |
SD | 19(8.4%) | 18(8.2%) | |
PD | 19(8.4%) | 52(23.5%) | |
ORR | 187(83.2%) | 151(68.3%) | <0.001 |
DCR | 206(91.6%) | 169(76.5%) | <0.001 |
Table 4
Cox Regression of Clinical Characteristics
Characteristics | Total(N) | HR(95% CI) | P value |
Age | 516 | | |
<=65 | 255 | Reference | |
>65 | 261 | 1.223 (0.916-1.635) | 0.172 |
Gender | 526 | | |
Female | 286 | Reference | |
Male | 249 | 1.070 (0.803-1.426) | 0.642 |
Smoker | 512 | | |
No | 75 | Reference | |
Yes | 446 | 0.894 (0.592-1.348) | 0.591 |
T stage | 523 | | |
T1 | 175 | Reference | |
T2 | 289 | 1.521 (1.068-2.166) | 0.020 |
T3&T4 | 68 | 3.066 (1.950-4.823) | <0.001 |
N stage | 510 | | |
N0 | 348 | Reference | |
N1 | 95 | 2.382 (1.695-3.346) | <0.001 |
N2&N3 | 76 | 2.968 (2.040-4.318) | <0.001 |
M stage | 377 | | |
M0 | 361 | Reference | |
M1 | 25 | 2.136 (1.248-3.653) | 0.006 |
Pathologic stage | 518 | | |
Stage I&Stage II | 417 | Reference | |
Stage III&Stage IV | 110 | 2.664 (1.960-3.621) | <0.001 |
CENPM | 526 | | |
Low | 268 | Reference | |
High | 267 | 1.665 (1.242-2.233) | <0.001 |
Expression Of Cenpm And Immune Infiltration
Next, we analyzed the immune infiltration data for the lung adenocarcinoma samples from the TCGA database. We found that the CENPM expression of CENPM was negatively correlated with the infiltration level of most immune cells (Figure 4A), such as macrophages (r = -0.120, p = 0.006, Figure 4B), B lymphocytes (r = -0.110, p = 0.015 Figure 4C), T lymphocytes (r = -0.096, p = 0.027, Figure 4D), and CD8 + T cells (r = -0.120, p = 0.004, Figure 4E). These results suggest an inhibitory status of the tumor immune microenvironment of lung adenocarcinomas with high CENPM expression.
We further analyzed the current possible correlation of immune checkpoint inhibition. High CENPM expression was positively correlated with PD-1 (PDCD1, r = 0.231, p < 0.001), and PD-L1 (CD274, r = 0.116, p = 0.007), but not correlated with TGFB1, and CTLA4 (Figure 4F, Table S2). These associations may serve as potential biomarkers for response with immune checkpoint inhibitors.
Cenpm-related Genes And Enrichment
We next analyzed the genes correlated with CENPM expression. The top 10 positively correlated genes and the top 10 negatively correlated genes were selected (Figure 5A). Further GO/KEGG enrichment of genes associated with CENPM expression identified the genes involved in the cell cycle-related pathways (Table S4), and genes associated with platinum resistance (Figure 5B).
CESA enrichment was further performed for genes associated with CENPM expression in all TCGA databases. We found that CENPM expression was enriched mainly in cyclins, and most genes were associated with poor prognosis in lung cancer. Furthermore, the CENPM expression was negatively correlated with T cell and B cell signaling pathways (Figure 5C).
Cenpm Gene Expression And Real-world Lung Adenocarcinoma
In our center, 20 patients with lung adenocarcinoma who submitted for gene expression profiling were analyzed. Among them, 6 patients had stage III and 14 patients had stage IV lung adenocarcinoma (Table S3). Four of these patients were treated with surgery and one with concurrent chemoradiotherapy. We found that CENPM expression was significantly higher in tumor tissues (n = 20) compared with adjacent non-cancerous tissues (n = 8) (2.767 vs. 1.337, p = 0.002; Figure 6A-B). We next analyzed the survival outcomes of these patient received chemotherapy. Patients with high expression of CENPM achieved poor PFS (12.5 months VS. 5.8 months, p=0.02, Figure 6C). However, OS was not reached and was not included in the analysis.