Analysis of TMEM88 mRNA expression
The analysis of TCGARNA-seq data using TIMER database showed that the expression of TMEM88mRNA in liver tumor tissues was higher than normal liver tissues (Fig. 1a). The same results are obtained when using GEPIA data analysis, and show significant differences in expression (P = 1.4e-15) (Fig. 1b). Continue to use GEPIA data to analyze the expression of TMEM88 in TCGA database in different pathological stages of liver tumors (stage I, II, stage III and IV). The results showed that the expression level of TMEM88 in each stage was significantly different from the normal liver tissues (P = 0.003) (Fig. 1c). The genetic variation of TMEM88 in different tumor types was analyzed using cBioPortal tool. The results showed that TMEM88 in liver tumor tissue affected the normal genetic function of liver tissue in the form of deep deletion, and the genetic variation of mutation or amplification was not detected in liver tumor tissue (Fig. 1d). In addition, we also used STRING tool to analyze the genes related to TMEM88 expression in liver tissue, and obtained four genes: DVL1, DVL2, DVL3 and VRK2 (Fig. 1e).
Furthermore, the potential relationship between TMEM88 and the expression of these genes in liver tissue was analyzed by using TIMER database. It was found that there was no significant relationship with DVL1, but it was significant and negatively correlated with both DVL2,DVL3 and VRK2 (Fig. 2a). However, when using GEPIA to analyze the correlation between these genes and TMEM88 expression, it was found that only the negative correlation that was statistically significant is VRK2 (P = 0.005; R=-0.14), but not with DVL2 and DVL3 (P = 0.26 and P = 0.37 respectively) (Fig. 2b).
Correlation of TMEM88 expression and the level of immune cell infiltration in HCC
In this study, data from 371 HCC patients in the TIMER database were used to study the relationship between TMEM88 expression and the level of immune cell infiltration in HCC tissues (Fig. 3a). The expression of TMEM88 in HCC tissues was positively correlated with the infiltration level of CD8 + T cells (partial.cor = 0.004), but there was no statistical significance (P = 9.4e-01). It was negatively correlated with Tumor Purity (partial.cor=-0.178; P = 9.04e-04), B cell (partial.cor=-0.207; P = 1.07e-04), Macrophage (partial.cor=-0.2; P = 1.97e-04) and Neutrophils (partial.cor=-0.145; P = 7.08e-03) infiltration. Although it was negatively correlated with the infiltration level of CD4 + T cells (partial.cor=-0.065), and Dendritic Cells (partial.cor=-0.05), there was not statistical significance (P = 2.27e-01 and P = 3.55e-01, respectively).
Furthermore, we found that there was a statistically significant correlation between the expression of TMEM88 in HCC and the survival outcome of patients (P < 0.005). Higher expression of TMEM88 in HCC patients was associated with better prognosis (Fig. 3b). There was no significant correlation between the expression levels of B cells (P = 0.489), CD8 + cells (P = 0.551), CD4 + cells (P = 0.454), Macrophages (P = 0.153), Neutrophils (P = 0.2), and Dendritic cells (P = 0.288) and the survival outcome of patients.
The TIMER database was used to analyze the potential correlation between the expression level of TMEM88 gene and the level of fibroblast infiltration in different tumor tissues (Fig. 3c). The results showed that MCPCOUNTER infiltration (P = 7.39e-03), XCELL infiltration (P = 1.27e-09) and TIDE infiltration (1.78e-05) of TMEM88 in liver tumor cells were positively correlated and statistically significant, while EPIC infiltration was almost zero (Fig. 3d).
Expression of TMEM88 in liver cancer tissue
Next, we used the Kaplan-Meier plotter database to analyze the prognostic value of TMEM88 expression in various human cancers. High TMEM88 expression was associated with better OS, PFS, RFS and DSS in the HCC cohort (n = 364, n = 370, n = 316, and n = 362 respectively). The low expression of TMEM88 was associated with poor prognosis of HCC overall survival rate (OS): HR = 0.58, 95%CI = 0.41 to 0.82, P < 0.005; progression-free survival (PFS): HR = 0.59, 95%CI = 0.44 to 0.8, P < 0.005; recurrence-free survival (RFS): HR = 0.69, 95%CI = 0.49 to 0.96, P = 0.026; disease-specific survival (DSS): HR = 0.49, 95%CI = 0.31 to 0.78, P = 0.49. Although the expression of TMEM88 mRNA in HCC is higher than that in corresponding normal tissues, high expression of TMEM88 is associated with a better prognosis of HCC (Fig. 4a).
The UALCAN tool was used to analyze the difference of TMEM88 phosphorylation level in different clinicopathological factors in normal liver tissue and liver tumor tissue. The results showed that the level of TMEM88 phosphorylation in normal liver tissue was significantly different from that in liver tumor tissue (Fig. 4b). In the analysis expression of TMEM88 in liver cancer tissue based on individual cancer stages, a total of 50 normal liver tissues, 168 Stage I liver cancer tissues, 84 Stage II liver cancer tissues, 82 Stage III liver cancer tissues and 6 Stage IV liver cancer tissues were collected. The results showed that there were significant differences between normal and Stage I (P = 2.33e-15), normal and Stage II (P = 1.76e-05), normal and Stage III (P = 3.81e-05), normal and Stage IV (P = 1.2e-02), and between Stage I and Stage II (P = 4.98e-02). However, there was no significant difference between Stage I and Stage III (P = 1.71e-01), Stage I and Stage IV (P = 4.39e-01), Stage II and Stage III (P = 7.55e-01), Stage II and Stage IV (P = 8.99e-01), and between Stage III and Stage IV (P = 8.28e-01).
TMEM88 expression in liver cancer tissue was analyzed based on patient race. A total of 50 normal liver tissue samples, 177 Caucasian patient tissue samples, 17 African American patient tissue samples and 157 Asian patient tissue samples were included. The results showed that there were significant differences between the normal liver tissues and Caucasian patient tissues (P = 2.2e-12), normal liver tissues and African American patient tissues (P = 1.24e-02), normal liver tissues and Asian patient tissues (P = 2.61e-10), and Caucasian and Asian patient tissues (P = 7.86e-03). There was no significant difference between the Caucasian and African American patient tissues (P = 5.77e-01) or African American and Asian patient tissues (P = 5.26e-01).
Analysis of the expression of TMEM88 in liver cancer tissue based on patient sex involved 50 normal liver tissue samples, 245 male patient tissue samples and 117 female patient tissue samples. The results showed that there were significant differences between the normal liver tissues and male patient tissues (P = 3.66e-15) and normal liver tissues and female patient tissues (P = 3.03e-07). However, there was no significant difference between the male and female patient tissues (P = 6.8e-01).
In the analysis of TMEM88 expression in liver cancer tissue based on patient age, a total of 50 normal liver tissue samples, 27 tissue samples from patients aged 21 to 40 years, 140 tissue samples from patients aged 41 to 60 years, 181 tissue samples from patients aged 61 to 80 years, and 10 tissue samples from patients aged 81 to 100 years were included. The results showed that there were significant differences between the normal liver tissues and those of patients of 21 to 40 years old (P = 1.46e-03), normal liver tissues and those of patients of 41 to 60 years old (P = 5.73e-11), normal liver tissues and those of patients of 61 to 80 years old (P = 9.67e-12), normal liver tissues and those of patients of 81 to 100 years old (P = 9.12e-02), tissues of patients of 21 to 40 and 41 to 60 years old (P = 1.78e-01), tissues of patients of 21 to 40 and 61 to 80 years old (P = 2.77e-01), and tissues of patients of 21 to 40 and 81 to 100 years old (P = 6.1e-01). There were no significant differences between the tissues of patients of 41 to 60 and 61 to 80 years old (P = 2.24e-01), tissues of patients of 41 to 60 and 81 to 100 years old (P = 3.09e-01), or tissues of patients of 61 to 80 and 81 to 100 years old (P = 3.88e-01).
Analysis of the expression of TMEM88 in liver cancer tissue based on patient weight involved 50 normal liver tissue samples, 154 tissue samples from patients with a normal weight (BMI greater than or equal to 18.5 and less than 25), 88 tissue samples from patients with an extreme weight (BMI greater than or equal to 25 and less than 30), 57 tissue samples from obese patients (BMI greater than or equal to 30 and less than 40) and 11 tissue samples from extremely obese patients (BMI greater than 40). The results showed that there were significant differences between the normal liver tissues and those from normal-weight patients (P = 5.38e-12), normal liver tissues and those from overweight patients (P = 5.24e-07), normal liver tissues and those from obese patients (P = 7.06e-07), normal liver tissues and those from extremely obese patients (P = 1.91e-02) as well as between the tissues from normal-weight and overweight patients (P = 5.2e-01), normal-weight and obese patients (P = 4.27e-01), normal-weight and extremely obese patients (P = 6.94e-01), and overweight and obese patients (P = 9.47e-01). There were no significant differences between the tissues from overweight and extremely obese patients (P = 9.85e-01) or obese and extremely obese patients (P = 9.85e-01).
TMEM88 expression in liver cancer tissue was analyzed based on tumor grade. Fifty normal liver tissue samples, 54 grade 1 patient tissue samples (well differentiated, low grade), 173 grade 2 patient tissue samples (moderately differentiated, intermediate grade), 118 grade 3 patient tissue samples (poorly differentiated, high grade) and 12 grade 4 patient tissue samples (undifferentiated, high grade) were assessed. The results showed that there were significant differences between the normal liver tissues and grade 1 patient tissues (P = 2.81e-10), normal liver tissues and grade 2 patient tissues (P = 1.03e-12), normal liver tissues and grade 3 patient tissues (P = 5.09e-07), grade 1 and grade 3 patient tissues (P = 2.18-02), grade 1 and grade 4 patient tissues (P = 9.20e-03), and grade 2 and grade 4 patient tissues (grade 9. 34 − 03). There were no significant differences between the normal liver tissues and grade 4 patient tissues (P = 1.26e-01), grade 1 and grade 2 patient tissues (P = 3.04e-01), grade 2 and grade 3 patient tissues (P = 9.03e-02), or grade 3 and grade 4 patient tissues (P = 9.91e-02).
Nodal metastasis status was also considered in the analysis of TMEM88 expression in liver cancer tissue. In this analysis, 50 normal liver tissue samples, 252 N0 patient tissue samples (no regional lymph node metastasis) and 4 N1 patient tissue samples (metastases in 1 to 3 axillary lymph nodes) were assessed. The results showed that there was a significant difference between the normal liver tissues and N0 patient tissues (P = 1.83e-12). There were no significant differences between the normal liver tissues and N1 patient tissues (P = 4.3e-01) or N0 and N1 patient tissues (P = 9.34e-01).
The association between TP53 mutation status and TMEM88 expression in liver cancer tissue was evaluated based on 50 normal liver tissue samples, 105 tissue samples with TP53 mutations and 255 tissue samples without TP53 mutations. The results showed that there were significant differences between the normal liver tissues and those with TP53 mutations (P = 8.78e-04), normal liver tissues and those without TP53 mutations (P < 1e-12), and tissues with and without TP53 mutations (P = 1.45e-04).
Finally, analysis of TMEM88 expression in liver cancer tissue based on histological subtype involved 50 normal liver tissue samples, 361 tissue samples from HCC patients, 3 tissue samples from fibrolamellar hepatocellular carcinoma (FL-HCC) patients and 7 tissue samples from hepatocholangiocarcinoma (HCC-CC) patients. The results showed that there were significant differences between the normal liver tissues and those from HCC patients (P = 7.77e-16) and between the HCC and FL-HCC patient tissues (P = 1.15e-02). There were no significant differences between the normal liver tissues and FL-HCC patient tissues (P = 1.08e-01), normal liver tissues and mixed HCC-CC patient tissues (P = 1.79e-01), HCC and mixed HCC-CC patient tissues (P = 2.66e-01), and FL-HCC and mixed HCC-CC patient tissues (P = 2.07e-01).
Table 1
Correlation of TMEM88 mRNA expression and prognosis in HCC patients with different clinicopathological factors by Kaplan–Meier plotter.
Clinicopathological factors | Overall survival | Progression-free survival |
N | Hazard ratio | P value | N | Hazard ratio | P value |
Sex | | | | | | |
Female | 118 | 0.44 (0.24–0.79) | 0.005 | 120 | 0.63 (0.38–1.05) | 0.0715 |
Male | 246 | 0.58 (0.37–0.91) | 0.016 | 246 | 0.65 (0.45–0.93) | 0.0173 |
Stage | | | | | | |
1 | 170 | 0.77 (0.42–1.41) | 0.3954 | 170 | 0.58 (0.35–0.96) | 0.0313 |
2 | 83 | 0.59 (0.27–1.31) | 0.1886 | 84 | 0.65 (0.36–1.18) | 0.1561 |
1 + 2 | 253 | 0.72 (0.45–1.16) | 0.1788 | 254 | 0.57 (0.39–0.84) | 0.0036 |
3 | 83 | 0.44 (0.24–0.81) | 0.0064 | 83 | 0.9 (0.53–1.55) | 0.7122 |
2 + 3 | 166 | 0.49 (0.3–0.8) | 0.0032 | 167 | 0.72 (0.48–1.07) | 0.1024 |
4 | 4 | - | - | 5 | - | - |
3 + 4 | 87 | 0.5 (0.28–0.89) | 0.0158 | 88 | 0.81 (0.48–1.37) | 0.4313 |
Grade | | | | | | |
1 | 55 | 0.59 (0.23–1.52) | 0.2734 | 55 | 0.72 (0.33–1.57) | 0.4078 |
2 | 174 | 0.53 (0.32–0.89) | 0.015 | 175 | 0.61 (0.39–0.94) | 0.0234 |
3 | 118 | 0.39 (0.21–0.75) | 0.003 | 119 | 0.55 (0.33–0.91) | 0.0172 |
4 | 12 | - | - | 12 | - | - |
AJCC_T | | | | | | |
1 | 180 | 0.79 (0.44–1.41) | 0.4186 | 180 | 0.62 (0.38–1.02) | 0.0557 |
2 | 90 | 0.47 (0.22-1) | 0.0466 | 92 | 0.62 (0.36–1.08) | 0.0887 |
3 | 78 | 0.47 (0.25–0.86) | 0.0123 | 78 | 0.75 (0.43–1.33) | 0.3288 |
4 | 13 | - | - | 13 | - | - |
Vascular invasion | | | | | | |
Micro | 90 | 0.46 (0.21-1) | 0.0452 | 91 | 0.68 (0.39–1.2) | 0.1818 |
None | 203 | 0.71 (0.43–1.19) | 0.194 | 204 | 0.68 (0.44–1.06) | 0.0888 |
Race | | | | | | |
White | 181 | 0.68 (0.43–1.07) | 0.0967 | 183 | 0.65 (0.44–0.97) | 0.0325 |
Asian | 155 | 0.43 (0.23–0.8) | 0.0059 | 155 | 0.6 (0.38–0.97) | 0.0346 |
Alcohol consumption | | | | | | |
Yes | 115 | 0.64 (0.34–1.22) | 0.1684 | 115 | 0.6 (0.35-1) | 0.0088 |
No | 202 | 0.5 (0.31–0.81) | 0.0037 | 204 | 0.66 (0.44–0.99) | 0.0456 |
Virus hepatitis | | | | | | |
Yes | 150 | 1.12 (0.58–2.16) | 0.7337 | 152 | 0.63 (0.39-1) | 0.0459 |
No | 167 | 0.39 (0.24–0.63) | 7.5e-05 | 167 | 0.63 (0.41–0.98) | 0.0369 |
Bold values indicate P < 0.05.
Correlation between TMEM88 expression and the clinical characteristics of HCC patients
Next, we investigated the relationship between TMEM88 expression and different clinical characteristics of HCC using the Kaplan–Meier plotter database, and the results are shown in Table 1. Low TMEM88 expression was correlated with both poorer OS and poorer PFS in men (OS: HR = 0.58, P = 0.002; PFS: HR = 0.65, P = 0.017), grade 2 patients (OS: HR = 0.53, P = 0.015; PFS: HR = 0.61, P = 0.023), grade 3 patients (OS: HR = 0.39, P = 0.003; PFS: HR = 0.55, P = 0.017), Asians (OS: HR = 0.43, P = 0.006; PFS: HR = 0.6, P = 0.035), and those who did not consume alcohol (OS: HR = 0.5, P = 0.004; PFS: HR = 0.66, P = 0.046). Low TMEM88 mRNA expression was correlated with worse OS in females (HR = 0.44, P = 0.005), stage 2 + 3 patients (HR = 0.49, P = 0.003), stage 3 patients (HR = 0.44, P = 0.006), stage 3 + 4 patients (HR = 0.5, P = 0.016), AJCC_T grade 2 patients (HR = 0.47, P = 0.047), AJCC_T grade 3 patients (HR = 0.47, P = 0.012) and patients with microvascular invasion (HR = 0.46, P = 0.045). Low TMEM88 mRNA expression correlated with worse PFS in stage 1 patients (HR = 0.58, P = 0.031), stage 1 + 2 patients (HR = 0.57, P = 0.004), white patients (HR = 0.65, P = 0.033), alcohol consumers (HR = 0.6, P = 0.009), patients with hepatitis virus (HR = 0.63, P = 0.046) and patients without hepatitis virus (HR = 0.63, P = 0.037). However, TMEM88 expression was not correlated with OS or PFS in stage 2 patients (OS: HR = 0.59, P = 0.189; PFS: HR = 0.65, P = 0.156), grade 1 patients (OS: HR = 0.59, P = 0.273; PFS: HR = 0.72, P = 0.408) or AJCC_T grade 1 patients (OS: HR = 0.79, P = 0.419; PFS: HR = 0.62, P = 0.056). These results demonstrate the prognostic significance of TMEM88 expression in patients with HCC based on their clinical characteristics, especially in the context of late-stage HCC.
Table 2
Expression correlation analysis between TMEM88 and marker genes of immune cells in GEPIA.
| | HCC |
Description | Gene markers | Tumor | Normal |
| | R | P | R | P |
Monocytes | CD86 | −0.1 | 0.054 | 0.57 | 1.8e-05 |
Neutrophils | CD11b | −0.23 | 9.4e-06 | 0.43 | 0.0019 |
CCR7 | 0.15 | 0.0035 | 0.49 | 0.00027 |
TAMs | CD68 | −0.1 | 0.051 | 0.48 | 0.00039 |
| IL-10 | −0.032 | 0.54 | 0.2 | 0.17 |
Th1 cells | IFN-γ (IFNG) | −0.14 | 0.0064 | 0.33 | 0.02 |
| STAT1 | −0.17 | 0.0012 | 0.45 | 0.00091 |
| T-bet (TBX21) | 0.17 | 0.00087 | 0.48 | 0.00046 |
| TNF-α (TNF) | −0.11 | 0.031 | 0.3 | 0.034 |
Th2 cells | STAT6 | 0.059 | 0.26 | 0.48 | 0.00047 |
Tregs | CCR8 | −0.12 | 0.018 | 0.13 | 0.18 |
| STAT5B | 0.012 | 0.81 | 0.26 | 0.18 |
| TGF-β (TGFB1) | −0.062 | 0.23 | 0.7 | 1.9e-08 |
Exhausted T cells | CTLA4 | −0.2 | 9.1e-05 | 0.43 | 0.0016 |
| PD-1 (PDCD1) | −0.13 | 0.016 | 0.46 | 0.00082 |
| TIM-3 (HAVCR2) | −0.18 | 4e-04 | 0.57 | 1.8e-05 |
Bold values indicate P < 0.05.
Correlation analysis between the mRNA levels of TMEM88 and markers of different subsets of immune cells
Next, we used GEPIA database to study the correlation between TMEM88 expression and type II T cell status according to the expression level of immune marker genes in HCC tissues. The immune cells analyzed in HCC tissues included monocytes, neutrophils, tumor associated macrophages (TAM), T helper cell 1 (Th1), T helper cell 2 (Th2), regulatory T (Tregs) and depleted T cells. GEPIA database analysis showed that the expression of TMEM88 in HCC was significantly correlated with the expression of marker genes in tumor-associated macrophages (TAM), neutrophils, T helper cells, Tregs, and depleted T cells (Table 2).
To be more specific, TMEM88 expression was close correlated with specific immune cell markers, such as monocyte marker CD86 (normal: r = 0.57, P = 1.8e-05) and CD11b (tumor: r=-0.23, P = 9.4e-06, normal: r = 0.43, P = 0.0019); neutrophil marker CCR7 (tumor: r = 0.15, P = 0.0035, normal: r = 0.49, P = 0.00027); and tumor associated macrophage marker CD68 (normal: r = 0.48, P = 0.00039). T helper cell 1 marker IFN- γ (tumor: r=-0.14, P = 0.0064, normal: r = 0.33, P = 0.02), STAT1 (tumour: r=-0.17, P = 0.0012, normal: r = 0.45, P = 0.00091), T-bet (tumour: r = 0.17, P = 0.00087, normal: r = 0.48, P = 0.00046) and TNF-α (tumour: r=-0.11, P = 0.031, normal: r = 0.3, P = 0.034); T helper cell 2 marker STAT6 (normal: r = 0.48, P = 0.00047); regulatory T cell marker CCR8 (tumor: r=-0.12, P = 0.018) and TGF- β (normal: r = 0.7, P = 1.9e-08); depleted T cell CTLA4 (tumor: r=-0.2, P = 9.1e-05, normal: r = 0.43, P = 0.0016), PD-1 (PDCD1) (tumor: r=-0.13, P = 0.016, normal: r = 0.46, P = 0.00082) and TIM-3 (HAVCR2) (tumor: r=-0.18, P = 4e-04; normal: r = 0.57, P = 1.8e-05).