Multiforme Glioblastoma, abbreviated to GBM, is a highly aggressive tumor of the brain and constitutes an important ongoing challenge in understanding its molecular underpinnings and finding preventive strategies (3). Epidemiological data from recent years paints an increasingly worrisome picture of increasing GBM incidence rates coupled with low survival rates (2). Thus, there exists a compelling rationale that calls for identifying early biomarkers for timely detection and possible prevention interventions.
An important part in GBM development is the role of Bone Morphogenetic Proteins (BMPs) which are members of the Transforming Growth Factor-β protein family. The proteins perform several roles such as: control CNS stem cell differentiation, apoptotic pathway, mitotic arrest and glioma-derived precursor cell differentiation (6, 37). Expression of dysregulated BMP protein has been implicated in oncogenic progression and modulation of tumor-suppressive pathways (39). However, the functional role of BMP proteins depends on the specific protein variant and tumor phenotype (7). In addition, oxygen tension has been identified as an influential factor in modulating BMP protein expression (12).
In high-grade glioma cells, hypoxic conditions have been shown to inhibit the BMP pathway through the HIF-1α protein (14). Understanding the intricate interplay between BMP protein expression patterns and HIF profiles in GBM patients remains a fascinating area for scientific inquiry. To address this, the current study aims to unravel the complex expression patterns of HIF factors and BMP pathway constituents within GBM tumor tissue, comparing them to non-neoplastic counterparts.
To analyze the enrichment pattern of genes from the BMP pathway in GBM cancer and normal count data, we employed the GSEA technique in this study. Interestingly, the GSEA results indicated a distinct enrichment pattern in the normal phenotype. Nevertheless, the observed enrichment failed to attain statistical significance, indicating that the differential enrichment of the BMP pathway in GBM cancer samples compared to normal samples is not substantial. Subsequently, we identified the top 14 core genes from the BMP pathway that drive the enrichment score of the GSEA clusters using Rank metric scores. These genes will undergo further analysis.
Moreover, a thorough examination was carried out on the variance in gene expression of HIF alpha transcription factors and the top 14 enriched genes from the BMP pathway in samples of GBM cancer. The analysis revealed that EPAS1, HIF3A, CHRDL1, NOG, BMP6, and AHSG genes did not exhibit a statistically significant difference between GBM cancer and normal tissue samples. In our previous research, HIF3A also showed no significant differential expression level in different types of TCGA cancer samples (32). However, the HIF1A gene showed significant upregulation in cancer samples, consistent with previous studies on the expression level of HIF1A in high-grade glioma cells under hypoxic conditions (14).
Further exploration of differentially expressed genes (DEGs) unveiled noteworthy findings. The majority of genes from the BMP pathway were significantly downregulated in GBM cancer samples compared to normal samples. Notable examples include SOSTDC1, GREM1, CHRD, SMAD7, PPM1A, BMPR2, MAPK1, ZFYVE16, GSK3B, and SMURF1 genes.
The correlation analysis by Pearson method has unveiled intriguing relationships between HIF alpha transcription factors and the top enriched genes from the BMP pathway in GBM cancer samples. Notably, the expression levels of HIF1A and EPAS1 correlated with PPM1A and BMPR2 genes while negatively correlated with SOSTDC1 and CHRDL1 genes. This negative correlation suggests a regulatory antagonist relationship between HIF1A and these genes within the BMP pathway, aligning with previous studies (14). While the investigation of the correlation between HIF1A and BMPR2 receptor expression in cancer cells and different tissue types has been limited in recent years, studies have highlighted that hypoxic conditions can downregulate the expression of the BMPR2 gene in lung cells of rats (28).
The MAPK1 gene also showed a moderate correlation with BMPR2 expression level. Furthermore, additional studies have underscored the participation of the Mitogen-activated protein kinase (MAPK) signaling pathway and MAPK-1 in the control of cellular death in neuronal cells subjected to stressful conditions. (29, 30, 31). The evaluation of HIF alpha factors and enriched genes from the BMP pathway in GBM cancer samples exhibits their potential as biomarkers with AUC values. Detecting GBM cancer at a stage is crucial, for improving treatment effectiveness and patient outcomes. To assess their potential, we conducted ROC analysis to evaluate both sensitivity and specificity of these biomarkers. Among the genes analyzed, seven displayed significant AUC values exceeding 0.90, including SOSTDC1, CHRD, SMAD7, PPM1A, MAPK1, SMURF1, and the HIF1A transcription factor. Previous investigations by other research groups revealed a significant relationship between SMAD7 and the TGF-beta signaling pathway in glioblastoma cells. These insights further support the importance of our findings (34, 35, 36, 38).
In addition, it is worth noting that a remarkable association was observed in survival analysis when examining the expression level of the SMURF1 gene and the survival rate of patients affected by GBM cancer. Interestingly, those individuals with a lower expression level of SMURF1 gene exhibited more favorable survival rates in comparison to those with a higher expression level. These outcomes closely align with previous investigations that have shed light on the oncogenic function of the SMURF1 gene during glioblastoma progression, underscoring its potential significance as a powerful prognostic biomarker for GBM cancer (33).
Furthermore, these findings distinctly emphasize the promise of these genes as both diagnostic and prognostic biomarkers for the timely detection of GBM cancer. The precise identification of these biomarkers has the potential to significantly contribute towards the design of effective therapeutic strategies and the enhancement of patient outcomes. Moreover, the intricately intertwined interaction between HIF alpha and the BMP pathway possesses tremendous potential in terms of identifying novel targets and developing innovative treatment approaches for individuals afflicted with GBM. Identification of new and specific biomarkers with help of bioinformatic techniques can help with the early detection of different types of cancer (40). However, the identified biomarkers should be evaluated by further experimental investigations in order to enter early-stage clinical phases.