Identification and validation of potential novel biomarkers for oral squamous cell carcinoma

Figures & data

Table 1. Clinical information of patients was included in this study

Characteristics	Oral cancer in situ (n = 3)	OSCC (n = 3)	Healthy controls (n = 3)
Gender
Male	0	1	2
Female	3	2	1
Age
≥50	3	2	0
<50	0	1	3
Cancer location
Tongue	0	1	0
Cheek	0	1	0
Gums	3	1	0
T stage
T1	0	0	0
T2	1	1	0
T3	2	2	0
T4	0	0	0
N stage
N0	2	2	0
N1	0	1	0
N2	1	0	0

Table 2. Differentially expressed proteins between oral carcinoma in situ and healthy control

	healthy controls VS oral carcinoma in situ
	AveExp. healthy controls	AveExp. oral carcinoma in situ	foldchange	P.Value
I309	4.59326	6.4722	7.85775	0.0249
GDF15	14.4222	16.3143	6.6718	0.0263
AXL	10.75932	13.47635	6.575194	0.0249
MMP3	11.3822	12.7522	5.3531	0.0178
CTACK	11.7833	12.5633	5.0141	0.01781

Table 3. Differentially expressed proteins between OSCC and healthy control

	healthy controls VS OSCC
Pro.	AveExp. healthy controls	AveExp. OSCC	foldchange	P.Value
MCSF	4.5433	7.6125	12.8621	0.0087
I309	4.5932	7.7886	12.6304	0.0092
MMP3	11.3822	13.4478	6.1642	0.0002
CTACK	11.7833	13.4456	5.7343	0.0168

Figure 1. Differentially expressed proteins in oral carcinoma in situ and OSCC. (a) Scatter plots of differentially expressed proteins between the oral carcinoma in situ group and the healthy control group. Red and Blue color indicated up- and down-regulation, respectively. (b) Scatter plots of differentially expressed proteins between the OSCC and the healthy control group. Red and Blue color indicated up- and down-regulation, respectively. (c) Comparison of the mean values of the significantly changed proteins between the cancer group and control normal group. The x-axis shows healthy control and cancer groups and y-axis shows mean values

Table 4. Correlation analysis between GDF15 and differentially expressed proteins

	Spearman Correlation	Sig.(2-tailed)
MCSF vs GDF15	0.601**	0.001
I309 vs GDF15	0.607**	0.001
MMP3 vs GDF15	0.603**	0.001
AXL vs GDF15	0.582**	0.001
CTACK vs GDF15	0.526**	0.005

**.Correlation is significant at the 0.01 level (2-tailed).

Table 5. The results of diagnostic efficiency analysis

Group	AUC	Youden’s index	Sensitivity	Specificity	P.Value	95% CI
GDF15	0.957	15.9	0.778	0.999	0.000	0.886–1.028
MMP3	0.969	12.77	0.833	0.999	0.000	0.913–1.025
AXL	0.914	12.15	0.778	0.889	0.001	0.807–1.021
MCSF	0.938	6.315	0.833	0.889	0.000	0.852–1.025
I309	0.951	6.59	0.833	0.889	0.000	0.872–1.029
CTACK	0.907	12.53	0.778	0.999	0.010	0.798–1.017

Figure 2. ROC curve analyses. The x-axis indicated specificity and y-axis indicated sensitivity. (a) GDF15. (b) MCSF. (c) I309. (d) MMP3. (f) CTACK. (e) AXL. (g) The combination of these six proteins

Figure 3. Kaplan-Meier analysis in TCGA dataset. (a) Risk score distribution, survival overview, and hierarchical clustering for OSCC patients in TCGA cohort. (b) The overall survival curves of the GDF15 high expression and low expression groups in TCGA cohort

Figure 4. Correlation analysis in TCGA dataset. (a) MMP3. (b) AXL. (c) MCSF. (d) I309. (e) CTACK. The X axis represents the expression distribution of GDF15, and the Y axis ordinate represents the expression distribution of MMP3, AXL, MCSF (CSF), I309 and CTACK, respectively. The density curve on the upper side is the distribution trend of GDF15, and the density curve on the right represents the distribution trend of MMP3, AXL, MCSF (CSF), I309 (CCL1) and CTACK (CCL27), respectively. (f) Correlation analysis between each gene and GDF15. Red represents a positive correlation, blue represents a negative correlation, the darker the color is, the stronger the correlation between the two, *P < 0.05, **P < 0.01

Table 6. The mean optical density of the immunohistochemical staining of cells

	Mean optical density (AOD) ($\bar{x}$ ± S)
Group	HSC-3	HaCaT	P.Value
GDF15	0.4696 ± 0.0519	0.3345 ± 0.0454	<0.0001
MMP3	0.3911 ± 0.0385	0.2980 ± 0.0138	< 0.0001
AXL	0.5154 ± 0.0818	0.2762 ± 0.0102	< 0.0001
MCSF	0.4603 ± 0.1116	0.3245 ± 0.0405	0.0039
I309	0.2910 ± 0.0457	0.2910 ± 0.0458	< 0.0001
CTACK	0.4391 ± 0.0348	0.3247 ± 0.0490	< 0.0001

Figure 5. Immunofluorescence staining of key proteins in HSC-3 and HaCaT cells. The cytoplasm of HSC-3 cells was stained with green fluorescence, and the cytoplasm of HaCaT cells was stained with red fluorescence. The nuclei were stained blue with DAPI

Figure 6. Immunohistochemical analysis of key proteins in HSC-3 and HaCaT cells

Figure 7. Western blot of key proteins in HSC-3 and HaCaT cells. The expression level of GDF15, MCSF, I309, MMP3, CTACK and AXL was detected by western blotting and its quantification in HSC-3 and HaCaT cells

Table 7. The mean optical density of the immunohistochemical staining of tissues

	Mean optical density ($\bar{x}$ ± S)
Group	OSCC tissue	Normal tissue	P.Value
GDF15	0.4657 ± 0.0291	0.3381 ± 0.0130	< 0.0001
MMP3	0.3661 ± 0.1076	0.2845 ± 0.0334	0.0433
AXL	0.4190 ± 0.0669	0.3523 ± 0.0254	0.0127
MCSF	0.6044 ± 0.0526	0.3522 ± 0.0255	< 0.0001
I309	0.4656 ± 0.0634	0.3459 ± 0.0282	< 0.0001
CTACK	0.3641 ± 0.025	0.3035 ± 0.0273	< 0.0001

Figure 8. In vivo tumorigenesis assay and histological evaluation. (a) In vivo tumorigenesis assay: (a) Tumor formation of nude mouse in experimental group. (b) The general view of tissue sampling in nude mouse. (c) The general view of tissue sampling in control group. (B) Histological evaluation: (a) HE staining of OSCC tumor tissues in nude mouse. (b) HE staining of normal tissues in control group. (n = 6)

Figure 9. Immunohistochemical analysis of key proteins in OSCC nude mouse

Figure 10. Western blot of key proteins in OSCC nude mouse. The expression level of GDF15, MCSF, I309, MMP3, CTACK and AXL was detected by western blotting and its quantification in OSCC nude mouse tumor tissue and normal tissue

Data availability statement

The data used and analyzed during the current study available from the corresponding author on reasonable request.