Expression of PLOD gene in pan-cancer
The expression levels of the PLOD family members were examined in all 18 cancer types in the TCGA database to discern the intrinsic expression pattern of PLOD (Fig. 1). Our results indicated that PLOD1 exhibited elevated expression in BLCA, BRCA, COAD, ESCA, GBM, HNSC, KIRC, KIRP, LIHC, LUAD, LUSC, PRAD, READ, STAD, THCA, and UCEC, while it was expressed at lower levels in CHOL and KICH (Fig. 1A). PLOD2 demonstrated higher expression in BLCA, BRCA, ESCA, GBM, HNSC, KIRC, KIRP, LIHC, LUAD, LUSC, STAD, THCA, and UCEC, whereas it had reduced expression in CHOL, COAD, KICH, PRAD, and READ (Fig. 1B). Notably, PLOD3 was found to have increased expression across all 18 cancer types (Fig. 1C). Upon further analysis of PLOD gene family expression within these cancers, PLOD1 and PLOD3 were observed to have higher expression, while PLOD2 showed lower expression (Fig. 2A). Additionally, the results indicated that PLOD1, PLOD2, and PLOD3 were most highly expressed in GBM (Fig. 2B). Moreover, correlations among various PLOD expression levels were observed (Fig. 3B). Based on the Spearman correlation test, among all pairwise gene correlations, a positive correlation was found between PLOD1 and PLOD2 (r = 0.31, P < 0001), with the strongest correlation observed between PLOD1 and PLOD3 (r = 0.47, P < 0001), suggesting potential shared features or functions. These findings revealed the presence of intrinsic disparities in PLOD gene family expression across various tumor types, indicating the importance of considering each PLOD family member as a distinct entity for investigation.
Correlation of PLODs gene expression with patient OS
Kaplan-Meier survival analyses revealed the association between PLOD gene expression and OS in patients with various cancer types. Elevated PLOD1 expression levels were associated with diminished survival in patients with ACC, BLCA, GBM, KICH, KIRC, KIRP, LGG, LIHC, MESO, and SARC (Fig. 3A). Similarly, heightened PLOD2 expression levels correlated with reduced survival in patients with BLCA, CESC, HNSC, KICH, KIRC, LGG, LIHC, MESO, PAAD, STAD, and UCS, while acting as a protective factor in SKCM (Fig. 3B). Increased expression of PLOD3 was also linked to poorer survival in CESC, COAD, HNSC, LGG, MESO, THYM, and UVM patients (Fig. 3C). All PLOD genes exhibited significant associations with OS in LGG patients (Fig. 3), with increased expression of PLOD1, PLOD2, and PLOD3 corresponding to heightened survival risk. The results demonstrated that PLOD1, PLOD2, and PLOD3 were notably correlated with patient survival rates (Table 1).
Subsequently, we explored the prognosis risk of the PLOD gene family through COX regression analysis. The findings suggested that PLOD genes were predominantly upregulated across the cancers evaluated, indicating that increased expression of PLOD genes was generally associated with adverse prognosis (Fig. 4). For instance, PLOD1, PLOD2, and PLOD3 were identified as high-risk factors in KICH patients, while PLOD2 served as a low-risk factor in SKCM patients. These findings suggested that PLOD genes may exhibit varying roles in the prognostic value across different cancers.
PLOD genes are correlated with tumor microenvironment and immune response in cancer
In human tumors, 6 types of immune infiltration have been identified, ranging from tumor-promoting to tumor-suppressing [15], termed C1 (wound healing), C2 (IFN-γ dominant), C3 (inflammatory), C4 (lymphocyte depleted), C5 (immunological quiet), and C6 (TGF-β dominant). There was a significant association between PLOD1/2/3 expression and these immune subtypes (p < 0.001) (Fig. 5A). Interestingly, PLOD1/2/3 showed low expression in C5, while exhibiting high expression in C1, C2, C3, C4, and C6. Both C1 and C6 subtypes, which are associated with poor prognosis, had elevated expression of PLOD1/2/3, suggesting their tumor-promoting roles. Moreover, PLOD1 and PLOD3 had significantly lower expression in C3 and C5 compared to C1, C2, and C4, implicating that higher gene expression is associated with deleterious immune profiles and further supports the notion of PLODs playing a fundamental role in cancer promotion.
Given that PLOD proteins can be secreted by fibroblasts, macrophages, and tumor cells in the TME, modulating tumor growth and metastasis [2, 16], we investigated the association between PLOD expression and mesenchymal cell levels. Figures 5B and 5C depict that PLOD expression is predominantly positively correlated with both immune and stromal scores, reflecting the high content of immune and stromal cells in 33 TCGA cancer types. Notably, a prominent positive association between immune scores, stromal scores, and PLOD expression was observed in LGG[11], aligning with prior studies that have reported high PLOD family expression to be linked with poor prognosis and elevated immune cell infiltration (Fig. 5B-C). Additionally, a significant association between stromal scores and PLOD expression was evident in DLBC, LGG, PCPG, and THYM (Fig. 5C). These findings suggest that PLODs exhibit differential capacities in modulating the immune microenvironment across various cancers.
PLOD genes are correlated with cancer cell sensitivity to chemotherapy and tumor stemness
PLOD genes have been implicated in both tumor stemness and cancer cell sensitivity to chemotherapy. We examined the relationship between tumor stemness scores and PLOD expression utilizing DNA stemness scores derived from DNA methylation patterns (DNAss) and RNA stemness scores based on mRNA expression levels (RNAss). Notably, PLOD expression was predominantly positively correlated with DNAss and inversely correlated with RNAss (Fig. 6A, B). These varying correlations suggested that RNAss and DNAss might delineate distinct cancer cell subsets, characterized by disparate properties or stemness signatures. Drug resistance remains a formidable challenge in clinical oncology and a key determinant of patient outcomes [17]. To elucidate the role of PLOD in drug sensitivity, we conducted an integrated analysis of drug sensitivity data and PLOD gene expression. Data from CellMiner™ revealed strong associations between PLOD expression and drug response, with specific genes influencing the directionality of these relationships (Fig. 6C). PLOD1 and PLOD3 expression were significantly inversely correlated with the sensitivity to Amonafide and Pyrazoloacridine, while PLOD2 exhibited a positive correlation with sensitivity to Lenvatinib. A positive correlation signifies that elevated gene expression aligns with favorable drug responses.
PLOD gene family in gastric cancer
To our knowledge, PLOD genes have been partially investigated in gastric cancer, with all members exhibiting tumor-promoting effects [3, 7, 18]. PLOD expression in tumors may be modulated by a potential alternative mechanism involving differential expression by various cell types within the TME. Although none of the patient samples fell under the C5 immune subtype in gastric cancer, all three PLOD genes exhibited significant associations with immune infiltration types (Fig. 7A). We further investigated the correlations between PLOD expression and gastric cancer grades, revealing no significant differences (Fig. 7B). Additionally, we assessed the relationships between PLOD expression and gastric cancer stemness scores. PLOD1 and PLOD2 expression showed significant negative correlations with the RNAss stemness score in gastric cancer (STAD), while PLOD3 displayed a significant positive correlation (Fig. 7C). Moreover, PLOD2 expression was negatively correlated with DNAss stemness scores, whereas PLOD1 and PLOD3 were positively correlated (Fig. 7C). In STAD patients, PLOD1 expression was positively associated with stromal scores and ESTIMATE scores (Fig. 7C). PLOD2 expression correlated positively with stromal, immune, and ESTIMATE scores in STAD (Fig. 7C), while PLOD3 expression showed negative correlations (Fig. 7C). These findings indicate that higher PLOD expression corresponds to increased tumor stemness indices, suggesting enhanced tumor stem cell activity and diminished tumor differentiation.