High TYROBP expression predicted better OS
To explore the expression level of TYROBP in osteosarcoma, the expression data from GSE42352 was analyzed and a significantly increased expression of TYROBP was observed in the osteosarcoma group compared with the mesenchymal stem cell group (P <0.001) (Figure 1A). Besides, TYROBP was highly expressed in osteosarcoma than that in osteoblast (P <0.001) (Figure 1B). To investigate the clinical benefits of TYROBP, we performed a Kaplan-Meier plotter analysis of the GDC TARGET-osteosarcoma dataset and found that osteosarcoma patients with high TYROBP expression exhibited better OS (Hazard ratio [HR] =0.14, P <0.001) (Figure 1C). Next, the ROC curve was used to demonstrate its value in distinguishing the survival status of osteosarcoma. As AUC was 0.674, TYROBP showed notably high specificity and sensitivity for discriminating the survival status (Figure 1D). The distribution of TYROBP expression, survival status of osteosarcoma patients, and expression profiles of TYROBP were presented in Figure 1E. Further, the relationship between TYROBP expression and OS was validated using GSE21257 data. Consistent with the TARGET results, the up-regulated TYROBP expression led to a favorable OS in osteosarcoma patients (HR =0.27, P <0.01) (Figure 2A). TYROBP expression presented potential prognostic value as the ROC curve showed that the AUC of TYROBP expression for predicting survival status was 0.625 (Figure 2B).
TYROBP was an independent prognostic factor in osteosarcoma
To determine the independent prognostic significance of TYROBP for OS in osteosarcoma, Cox regression analysis was conducted. Univariate regression analysis showed that non-metastasis (HR =0.225, P <0.001) and high TYROBP expression (HR =0.584, P <0.001) were significantly related to the prolonged OS time. However, age, gender, leg/foot, and pelvis had no remarkable relationship with the clinical outcomes. In multivariate regression analysis, non-metastasis (HR =0.245, P =0.003) and high TYROBP expression (HR =0.629, P =0.005) were still independent factors for favorable OS in patients with osteosarcoma (Table 1).
Table 1
Cox regression analysis of TYROBP expression and overall survival for patients with osteosarcoma
|
Characteristics
|
Univariate analysis
|
Multivariate analysis
|
HR (95 % CI)
|
P-value
|
HR (95 % CI)
|
P-value
|
Age
|
1.005 (0.921-1.097)
|
0.908
|
1.057 (0.957-1.167)
|
0.272
|
Gender (male vs. female)
|
0.775 (0.346-1.737)
|
0.535
|
0.699 (0.276-1.773)
|
0.451
|
Metastasis status (M0 vs. M1)
|
0.225 (0.100-0.505)
|
<0.001
|
0.245 (0.098-0.612)
|
0.003
|
Leg/foot
|
0.963 (0.129-7.187)
|
0.971
|
1.585 (0.178-14.155)
|
0.680
|
Pelvis
|
4.606 (0.413-51.309)
|
0.214
|
7.989 (0.590-108.217)
|
0.118
|
TYROBP
|
0.584 (0.436-0.781)
|
<0.001
|
0.629 (0.456-0.868)
|
0.005
|
Abbreviations: HR, hazard ratio; CI, confidence interval; M0, no metastasis; M1, metastasis. |
To better predict the prognosis of osteosarcoma patients, a nomogram based on the Cox regression analysis results was constructed and a calibration curve was plotted to evaluate the efficiency of the nomogram. Two statistically significant prognostic factors metastasis status and TYROBP expression were included in the model to predict the OS, which had a C-index of 0.774 (Figure 3A). The calibration curve presented a desirable prediction of the nomogram for the 3-, and 5- year survival probability (Figure 3B).
Prognostic significance of TYROBP in the osteosarcoma subgroups
Subsequently, we examined the prognostic value of TYROBP for OS in certain clinicopathological subgroups. Multivariate Cox regression analysis was conducted in specific subgroups. As shown in Table 2, TYROBP was an independent prognostic factor for OS in patients of male sex (HR =0.528, P =0.016), age below 18 years (HR =0.647, P =0.012), metastasis (HR =0.566, P =0.043), and tumor site at leg/foot (HR =0.585, P =0.002). Likewise, the forest plot illustrated the independent prognostic value of TYROBP in osteosarcoma with restricted characteristics using multivariate Cox regression results (Figure 4). The subgroup analyses for tumor site (arm/hand, and pelvis) were not performed due to few samples. The Kaplan-Meier plotter analysis for OS was performed in these four subgroups: male, age below 18 years, metastasis, and tumor site leg/foot (all P <0.05) (Figure 5A-5D). All the results showed significantly longer OS in the high TYROBP expression groups.
Table 2
Prognostic performance of TYROBP on overall survival in osteosarcoma patient subgroups by multivariate Cox regression analysis
|
Characteristics
|
N (%)
|
Hazard ratio (95%CI)
|
P-value
|
Gender
|
|
|
|
Female
|
34 (44.7)
|
0.76 (0.436-1.048)
|
0.08
|
Male
|
42 (55.3)
|
0.528 (0.315-0.887)
|
0.016
|
Age
|
|
|
|
<18
|
61 (80.3)
|
0.647 (0.460-0.910)
|
0.012
|
≥18
|
15 (19.7)
|
0.437 (0.076-2.514)
|
0.437
|
Metastasis status
|
|
|
|
M1
|
18 (23.7)
|
0.566 (0.327-0.981)
|
0.043
|
M0
|
58 (76.3)
|
0.663 (0.387-1.136)
|
0.134
|
Tumor site
|
|
|
|
Leg/foot
|
70 (92.1)
|
0.585 (0.418-0.818)
|
0.002
|
Correlation between TYROBP expression and clinical characteristics
Since high TYROBP mRNA expression was closely associated with favorable OS, we next explored the clinicopathological factors that could affect its mRNA expression. As shown in Table 3, patients in the high TYROBP expression group manifested a higher proportion of leg/foot primary tumor sites than those in the low TYROBP expression group (P <0.05). Whereas, there was no significant difference in the distribution of age, gender, or metastasis between the two groups (all P >0.05).
Then, we examined the TYROBP expression in osteosarcoma patients with various clinicopathological parameters. Patients in different groups of age, gender, and metastasis shared similar TYROBP mRNA expression levels (all P >0.05) (Figure 6A-6C), while leg/foot exhibited the highest TYROBP mRNA expression compared with other primary tumor sites with a statistical difference (P <0.01) (Figure 6D).
Table 3
Relationship between TYROBP levels and clinicopathological parameters of osteosarcoma in TARGET
|
Variables
|
TYBOBP
|
|
χ2
|
P-value
|
Low (%)
|
High (%)
|
|
|
Age (years)
|
|
|
0.083
|
0.773
|
<18
|
31 (50.8)
|
30 (49.2%)
|
|
|
≥18
|
7 (46.7)
|
8 (53.3%)
|
|
|
Gender
|
|
|
0.213
|
0.645
|
Female
|
18 (52.9)
|
16 (47.1%)
|
|
|
Male
|
20 (47.6)
|
22 (52.4%)
|
|
|
Metastasis status
|
|
|
2.621
|
0.105
|
Metastasis
|
12 (66.7)
|
6 (33.3%)
|
|
|
Non-metastasis
|
26 (44.8)
|
32 (55.2%)
|
|
|
Primary tumor site
|
|
|
8.833
|
0.012
|
Arm/hand
|
4 (100.0)
|
0 (0.0%)
|
|
|
Leg/foot
|
32 (45.7)
|
38 (54.3%)
|
|
|
Pelvis
|
2 (100.0)
|
0 (0.0%)
|
|
|
TYROBP regulated the progression of osteosarcoma mainly through immune-related pathways
To elucidate the pathological function of TYROBP in osteosarcoma, we performed the GO annotation and KEGG pathway analyses of DEGs between high and low TYROBP expression groups. Through differential expression analysis, we obtained 175 DEGs including 166 upregulated and 9 downregulated genes as shown in the volcano plot (Figure 7A). The heat map presented the top 50 DEGs (Figure 7B). In terms of the cellular components, the DEGs were mainly enriched in a secretory vesicle, secretory granule, and vacuole (Figure 7C). The major molecular functions were signaling receptor binding, protein-containing complex binding, and amide binding (Figure 7D). For biological processes, they were mainly involved in defense response, immune effector process, and regulation of immune system process (Figure 7E). The KEGG pathways in which they mainly participated were cell adhesion molecules, cytokine-cytokine receptor interaction, osteoclast differentiation, antigen processing and presentation, natural killer cell-mediated cytotoxicity, NOD-like receptor signaling pathway, and chemokine signaling pathway, most of which are known to contribute to antitumor immunity. (Figure 7F).
Further, GSEA was conducted to reveal the underlying mechanism of TYROBP in osteosarcoma. The results identified that the biological pathways enriched in TYROBP high expression phenotype were lysosome, B cell receptor signaling pathway, natural killer cell-mediated cytotoxicity, Fc gamma R-mediated cytotoxicity, and antigen processing and presentation (Figure 8). Taken together, TYROBP might affect the OS of osteosarcoma patients by regulating the antitumor immune-related pathways.
Association of TYROBP with the immune cell infiltrates
We have demonstrated that TYROBP was mainly involved in immune-related pathways, and hence we attempted to explore the correlation between TYROBP and immune cell infiltration levels using the ESTIMATE algorithm. As shown in Figure 9A, the high TYROBP expression group had a significantly higher immune score, indicating a higher proportion of immune cell infiltrates (P <0.001). Besides, the stromal score in the high TYROBP expression group was not significantly different from that in the low TYROBP expression group (Figure 9B). Correlation analysis showed that TYROBP had a strong positive relation with immune score (P <0.001, r=0.87) (Figure 9C). In addition, patients with high immune scores had longer OS (P <0.01) (Figure 9D). These results confirmed that TYROBP might improve the clinical outcomes of osteosarcoma patients via positive regulation of the antitumor immunity.