PHLDA2 expression analysis
The expression of PHLDA2 in TCGA pan-cancer was investigated and it was discovered that the expression of PHLDA2 in tumor tissues was higher in comparison to normal tissues in 21 out of 33 tumors, including HCC (Figure 1A). Within the TCGA cohort, PHLDA2 expression exclusively in tumor tissues showed highest expression in ESCA and lowest expression in LAML (Figure 1B) while in normal tissues from the GTEx cohort, bone marrow had the highest expression, and blood had the lowest expression (Figure 1C). PHLDA2 expression was also high in tumor tissues in 11 tumor types, including HCC (Figure 2A-L) when compared to expression in normal tissues. It was also observed that PHLDA2 expression was relatively higher in worse tumor stages in ACC, BLCA, DLBC, KIRC, LIHC, LUAD, PAAD, and THCA ( Figure 2M-T).
Genetic alteration analysis of PHLDA2
Using data from cBioportal database, we evaluated the genetic modification of PHLDA2. We computed the association between PHLDA2 and copy number alteration (CNA) as well as methylation level in pan-cancer. The results showed a positive correlation between PHLDA2 expression and its CNA (Figure 3A-B) in pan-cancer and HCC and a negative correlation with its methylation level (Figure 3C-D) in pan-cancer and HCC, signifying that the genetic alteration of PHLDA2 has an impact on PHLDA2 mRNA expression.
The prognostic role of PHLDA2 in pan-cancer
Forest plots were employed for uniCox analysis of PHLDA2 in pan-cancer, including OS, DSS, DFI, and PFI, in order to further examine its prognostic role. Findings showed that PHLDA2 is a risk factor for OS in LGG, UVM, LIHC, KIRC, PAAD, LAML, GBM, LUAD, and ACC (Figure 4A), and for DSS, it is a risk factor in LGG, KIRC, UVM, LIHC, LUSC, PAAD, GBM, and BRCA (Figure 4B). PHLDA2 was found to be a risk factor in PRAD for DFI (Figure 4C) and high expression of PHLDA2 was predictive of longer PFI in LGG, KIRC, UVM, PAAD, LUSC, GBM, and SARC (Figure 4D).
Further analysis of OS using Kaplan-Meier demonstrated that high PHLDA2 expression indicates worse OS in 17 tumors including ACC, BRCA, CESC, COAD, GBM, HNSC, KIRC, LAML, LGG, LIHC (HCC), LUAD, LUSC, MESO, OV, PAAD, SKCM, and UVM (Sup-Figure 1). Moreover, multivariate cox regression analysis indicated that PHLDA2 acts as an independent prognostic factor in ACC, GBM, KIRC, LAML, LGG, LIHC, LUAD, PAAD, and UVM (Figure 5). Hence, these findings suggest that PHLDA2 is highly expressed in tumor tissues and indicates worse survival status for HCC patients and most tumor types.
Functional analysis of PHLDA2
We conducted a correlation analysis of PHLDA2 with all mRNAs to determine its possible functional pathways in HCC (Figure 6A-B). Using the correlation results, we performed GSEA of PHLDA2 in HCC based on GO, KEGG, and Reactome pathways. GO-based GSEA revealed that PHLDA2 was predominantly linked to malignant oncogenic pathways and immunoregulation-related pathways like myeloid leukocyte activation, regulation of leukocyte activation, and leukocyte activation involved in immune response (Figure 6C). KEGG pathway-based GSEA indicated that PHLDA2 was mainly related to IL−17 signaling pathway, Proteoglycans in cancer, TNF signaling pathway, and Neutrophil extracellular trap formation (Figure 6D). Reactome pathway-based GSEA suggested that PHLDA2 was mostly connected to Neutrophil degranulation, Signaling by Rho GTPases, and Innate Immune System (Figure 6E). Additionally, the GSVA results supported the positive association of PHLDA2 with malignant oncogenic pathways and immunoregulation-related pathways in HCC and pan-cancer, such as P53 PATHWAY, GLYCOLYSIS, TNFA SIGNALING VIA NFKB, and ANGIOGENESIS (Figure 7A-B). Overall, these findings strongly suggest that PHLDA2 is significantly linked to the tumor immune microenvironment in HCC.
PHLDA2 is associated with immunosuppressive microenvironment in HCC
By exploring the relationship between PHLDA2 and TME in HCC, we have found that PHLDA2 has a positive correlation with immune and stromal scores, and a negative correlation with tumor purity score in both HCC and pan-cancer (Figure 8A-E).
Seeking to further understand this association, we examined the link between PHLDA2 and immune cell infiltration in each cancer using data from the TIMER2. Our results show that PHLDA2 is positively correlated with the majority of immune cells, especially TAMs in HCC (Figure 9A). To confirm our findings, we utilized data from published studies (Figure 9B-C) and the ImmuCellAI database (Figure 9D-F), which revealed a positive correlation between PHLDA2 and TAMs and Tregs.
We have proved the association of PHLDA2 with immunosuppressive cells. We further assessed the potential relationship of PHLDA2 with immune-related genes. We found that PHLDA2 expression was positively related with immunosuppressive genes (Figure 10A), chemokines (Figure 10B), chemokine receptors (Figure 10C) and MHC genes (Figure 10D) in HCC and pan-cancer. PHLDA2 was also observed to be associated with immune checkpoints in HCC (Figure 10E). Tgfb1 signaling and wnt beta-catenin signaling were reported with immunosuppressive TMEs, we found that PHLDA2 has close relationship with these pathways (Figure 11A-B). In conclusion, our results demonstrated that PHLDA2 is a potential regulator of tumor immunosuppressive microenvironment.
Drug resistance analysis of PHLDA2
Furthermore, we examined the expression of PHLDA2 in conjunction with the IC50 of anti-cancer drugs based on GDSC database (Supplementary Table 2). Of the 192 anti-cancer drugs analyzed, 165 showed a positive association between their IC50 values and PHLDA2 expression levels, such as Daporinad, MIRA-1, Sabutoclax, AZD5991, Venetoclax, Leflunomide, Vorinostat, Zoledronate, and IWP-2 (Figure 12). This implies that individuals with elevated PHLDA2 expression may have a diminished response to these anti-cancer drugs.