1. Trem2 is highly expressed in HCC
Screening the expression of Trem2 in all primary tumors through the TIMER2.0 online database (http://timer.cistrome. org/)[13] showed that Trem2 was significantly up-regulated in 17 cancer types including HCC (1A), indicating the importance of Trem2 in tumorigenesis and development. The differential analysis and paired differential analysis of the mRNA expression of trem2 in the 371 HCC cases downloaded from the TCGA database showed that the transcription level of trem2 in the tumor tissue was significantly higher than that in the homologous adjacent normal liver tissue (1B, 1C). Although there are a few transcriptome data showing the relative down-regulation of trem2 in HCC tissues, this may be due to individual differences or heterogeneity of cell promiscuity. Therefore, we verified it in our patient samples. Excluding the pathogenic factors and background, through immunohistochemistry and immunofluorescence examination of 12 HCC patients, we found that the expression of trem2 in HCC was significantly increased (1D, 1E). Transcription examination of HCC tissue and distal tissues of each patient also showed a significant upregulation of Trem2 (data not shown).
2. Trem2 is an independent prognostic factor of HCC and has nothing to do with whether patients have an HBV background.
The COX test was performed on the HCC data of the TCGA cohort to evaluate the impact of various risk factors on the survival time of HCC patients. The results showed that univariate or multivariate cox regression analysis identified Trem2 as an independent prognostic factor affecting the overall survival of HCC patients (p = 0.01, 0.04, 2A, table 1). HCC patients often have an HBV background[14]. Therefore, we need to know whether Trem2 depends on the HBV background to participate in the regulation of HCC process. Wilcox test analysis showed that trem2 does not rely on HBV to participate in the regulation of HCC (P = 0.17, 2B), it means that trem2 may drive the development of HCC by not alone responding to the damage caused by viral DNA.
3. The high expression of Trem2 is related to the clinical stage and grade of HCC and indicates a poor prognosis
Tumor TNM staging and Grade classification are important references for clinical evaluation of tumor progression and prognosis[15]. We checked the internal relationship between trem2 and TNM staging and Grade in HCC through the R software (3.6.3). We found that Trem2 and T staging and Grade have a significant positive correlation (p = 2.1e-13, 1.5e-13, 3A). We then examined the relationship between the expression of trem2 and the prognosis of HCC patients through survival analysis. We found that high expression of trem2 predicts a worse prognosis for HCC patients (p = 0.0086, 3B).
4. Trem2 expression in different HCC mutation backgrounds
The detection of the whole transcriptome combined with the proteome identified TP53, AXIN1, CTNNB1, KEAP1 and RB1 as the genes with the highest mutation tendency in HCC [14]. TP53 is the most common mutant gene in HCC. Not surprisingly, HCC with TP53 mutation has a higher trem2 transcription tendency than WT HCC (P = 3.4E-06, 4A); according to the study of Gao et al.[14], Among patients in the CHCC-HBV cohort, the mutation frequency of AXIN1 was higher than that of the HBV cohort in the TCGA database. The expression of trem2 in HCC patients with AXIN1 mutations screened by TIMER 2.0 was not significantly different from that of the WT group (P = 0.18, 4C). Although the number of samples in the AXIN1 mutation group is small (n = 23/500), the no difference between the groups is more likely to be related to whether the HBV background has no significant effect on the expression of Trem2; CTNNB1 mutations rely on various metabolic pathways to drive HCC progression, such as Drug metabolism, glycolysis/gluconeogenesis and amino acid metabolism. Unlike the TP53 mutation, the expression of Trem2 was significantly down-regulated in patients with CTNNB1 mutation (p = 1.8e-5, 4A). This prompted us to explore the potential role of trem2 in regulating liver metabolism. We have used cluster Profiler R package confirmed through Go functional enrichment and KEGG (Kyoto Encyclopedia of Genes and Genomes) enrichment analysis that the expression of trem2 is indeed negatively correlated with pathways including drug metabolism and glycolysis/gluconeogenesis (4A), which reflects trem2 expression has a high compliance that is strongly negatively correlated with CTNNB1 mutation. In addition, trem2 is also most negatively related to retinol metabolism and chemical carcinogenesis (4C). The pathways for trem2 up-regulation include P53, FC gamma R and other pathways (4D); Keap1 and RB1 mutations also have no significant effect on the expression of trem2 (p = 0.49,0.24,4A). The volcano map shows that 2395 genes including trem2 are up-regulated and 463 genes are down-regulated in HCC in the TCGA database (4B).
5. The correlation between Trem2 in HCC and multiple immune checkpoints
TIMER2.0 was used to measure trem2 and PDL1 (CD274, Spearman's rho = 0.268, p = 1.57e-07), VEGFA (Spearman's rho = 0.152, p = 3.32e-03), CD276 (Spearman's rho = 0.488, p = 1.36e-23), TIGIT (Spearman's rho = 0.453, p = 3.87e-20), CTLA4 (Spearman's rho = 0.507, p = 1.23e-25), SIRPa (Spearman's rho = 0.342, p = 1.37e-11) and other immune checkpoints were positively correlated (5A-F). The currently approved clinical systemic immune strategies for the treatment of HCC are not very effective. Although current clinical trials show that when anti-PD-1 or anti-PD-L1 drugs are used in combination with anti-CTLA4, the response rate and response duration are satisfactory to some extent[16], we still need to explore new methods to enhance the immunosuppressive efficacy of HCC therapy. Therefore, our analysis shows that targeting trem2 or combined targeting may be a promising strategy in HCC.
6. GO and KEGG identified Trem2 is related to fatty acid metabolism
A large study showed that genes related to FA biosynthesis in HCC tissues were significantly up-regulated[17]. In fact, the generation of FA shows considerable potential in driving the progress of HCC. KEGG identified fatty acid metabolism (ES: -0.87) as the functional pathway with the highest negative correlation enrichment index for trem2. Among the identified fatty acid metabolism pathways, genes with high negative correlation ES with Trem2 mainly include ACSL (ACSL1, ACSL3, ACSL4, ACSL5 and ACSL6) family and ACAT family (ACAT1, ACAT2) (Table 2). The ACSL family exists in mammals and contributes to anabolic lipid biosynthesis and catabolic fatty acid oxidation. Most current research results indicate that the ACSL family has a cancer-promoting effect [18]. ACLY (ATP citrate lyase) is an enzyme involved in lipid synthesis. Multiple methods have proven that ACLY can be used as a prognostic marker of HCC [19]. Increased expression and phosphorylation of ACLY, ACSL3, and ACSL4 in HCC have been shown to activate lipid biosynthesis in HCC [14]. In HCC, Trem2 has a positive correlation with the expression of ACLY on the one hand (rho = 0.283, p = 3.04e-08, 6B), on the other hand, it has a negative correlation with the expression of ACSL1. Surprisingly, the expressions of ACSL3 and ACSL4 and trem2 were positively correlated in HCC after purity adjustment by TIMER 2.0 database, this is interesting. After further exploration, ADH1A is also a highly negatively correlated ES gene of trem2 enriched in the fatty acid metabolism pathway. It was previously found to be down-regulated in different subtypes of HBV-HCC and indicated a poor prognosis, and the worst prognosis S-Pf The expression of ADH1A in proliferative HBV-HCC was the lowest[14]. Trem2 transcription can significantly down-regulate the expression of ADH1A (6E, Table 2). HCCs with high ADH1A expression have fewer TP53 mutations, which in turn leads us to think about the internal connection between trem2, ADH1A and TP53 mutations. PPAR-a (rho=-0.126, p = 1.54e-02, 6A) and ACAT1 (rho=-0.202, p = 1.63e-04, 6D) have also been shown to have a significant negative correlation with trem2. Analyze the protein network of Trem2 by STRING (string-db.org). It is already known that Trem2 is mainly enriched with TYROBP and SYK (6F). Then we customized analysis of fatty acid pathway-related genes enriched in trem2 screened from KEGG in STRING. Although trem2 is currently unable to be cross-linked into the fatty acid metabolism grid due to the research blank, its enriched related genes have constructed a complete fatty acid metabolism grid. In general, for HCC RNA-Seq, the role of trem2 in fatty acid metabolism is contradictory. This may be due to the identity confounding of whole transcriptome sequencing, or it may be due to Trem2 foreshadowing huge cellular heterogeneity. Regardless, these results prompt us to explore the powerful potential of Trem2 in fatty acid metabolism and metabolism-driven HCC.
7.TREM2 may regulate the progression of HCC by promoting fibrosis
KEGG enrichment analysis proved that Trem2 has a positive correlation with fibrosis regulatory factors Col1A1/2 and MMP9 (7A)[20]. PPARα has been shown to improve fatty degeneration, inflammation and fibrosis of NAFLD, Therefore, Trem2 negatively regulates PPARa to affect liver lipid metabolism and fibrosis is one of the possible mechanisms[21]. Its indicates that Trem2 may be involved in regulating the occurrence of fibrosis. Fibrosis is often the common destination of most hepatitis B infections and NAFLD progression[22, 23]. Detect the expression of Trem2 and αSMA in human fibrosis (quantified by Ishak staging) and HCC with cirrhosis (7B-C). Trem2 showed that the expression level gradually increased with the progress of fibrosis and HCC, and it co-localized strongly with HSC of αSMA mark. This shows that Trem2 is a positive factor in promoting fibrosis. Subsequently, we induced fibrosis during CCL4, BDL, CDAHFD60 models to detected the high expression of Trem2 (7D-F).