Clear cell renal cell carcinoma (ccRCC) is the most common type of kidney cancer, representing 70-80% of all kidney cancers. While surgery is very effective for early-stage disease, advanced ccRCC is notoriously difficult to treat - it responds poorly to chemotherapy and radiation, and not all patients benefit from newer targeted or immunotherapy treatments.
Identifying reliable biomarkers - measurable biological signals that predict cancer behavior and treatment response - is critical for improving patient outcomes. Many biomarkers have been studied in kidney cancer, but few are currently used in routine clinical practice to guide decisions about which patients need aggressive treatment or which therapies they are most likely to benefit from.
The gene BBOX1 (gamma-butyrobetaine dioxygenase) is responsible for making L-carnitine, a molecule involved in fat metabolism and energy production. BBOX1 is highly expressed in normal kidney tissue and has been linked to cancer behavior in several other cancer types including breast, ovarian, and colorectal cancer. However, its role in kidney cancer had not been systematically investigated before this study.
This study analyzed BBOX1 expression in 857 kidney cancer patients from two independent cohorts - one from Hanyang University Hospital in South Korea and one from The Cancer Genome Atlas (TCGA) database - using a combination of immunohistochemistry, computational biology, and machine learning to understand BBOX1's clinical significance.
The study drew on data from two independent patient populations. The first was 203 ccRCC patients from Hanyang University Hospital (HYH) who underwent surgery between 2006 and 2017. The second was 533 ccRCC patients from The Cancer Genome Atlas (TCGA) with available RNA sequence data. Using two separate datasets allows findings to be discovered in one group and confirmed in another.
In the hospital cohort, BBOX1 protein expression was measured directly in tumor tissue using immunohistochemistry (IHC) on tissue microarray blocks. Each tumor sample was scored for both staining intensity (0-3) and the percentage of tumor cells staining positive (1-4 categories), and these were multiplied to give an immunoreactive score (IRS). Tumors were classified as low BBOX1 (IRS less than 1) or high BBOX1 (IRS 1 or above).
In the TCGA cohort, BBOX1 expression was measured at the RNA level, with patients divided into low and high BBOX1 groups using a cutoff derived from ROC curve analysis. Gene expression patterns, immune cell compositions (via CIBERSORT), and molecular pathways were then compared between the two groups.
A machine learning algorithm called gradient boosting machine (GBM) was used to assess how much BBOX1 adds to survival prediction beyond conventional clinical factors like tumor stage and grade. Drug sensitivity data from the Genomics of Drug Sensitivity in Cancer (GDSC) database were also analyzed to identify drugs that work better against tumors with low BBOX1 expression.
In the hospital cohort, low BBOX1 expression was significantly associated with higher histological grade - indicating more aggressive tumor cell appearance under the microscope - and with sarcomatoid change, a particularly dangerous form of tumor transformation associated with very poor prognosis (p=0.016 and p=0.019, respectively).
Survival analysis in both patient cohorts confirmed that low BBOX1 expression was linked to significantly worse outcomes. In the hospital cohort, patients with low BBOX1 had worse disease-specific survival (DSS, p=0.011) and overall survival (OS, p=0.003). The TCGA analysis confirmed these findings with even stronger statistical significance (all p values less than 0.001).
Multivariable analysis confirmed BBOX1 as an independent predictor of both DSS and OS - meaning its impact on survival was separate from other known risk factors such as tumor stage, histological grade, and lymphovascular invasion. For overall survival, low BBOX1 corresponded to a hazard ratio of 2.74, meaning patients with low BBOX1 had nearly triple the risk of dying during follow-up.
The machine learning model showed that adding BBOX1 expression to clinical factors significantly improved survival prediction accuracy. The area under the ROC curve increased from 0.993 to 0.999 when BBOX1 was included - a meaningful improvement demonstrating its added value over conventional clinical parameters alone.
Gene Set Enrichment Analysis (GSEA) using TCGA data identified seven biologically significant gene sets enriched in patients with low BBOX1 expression. These included pathways associated with: Mel-18 (a regulator of tumor cell proliferation and angiogenesis), p53 (a key tumor suppressor), and epithelial-mesenchymal transition (EMT) - a process by which cancer cells become more aggressive and better able to spread.
Additional enriched pathways included the KEGG cancer pathway (a broad catalog of cancer-related biological signals), an invasiveness signature, and the PTEN pathway - a critical tumor suppressor that regulates cell growth and is frequently disrupted in advanced kidney cancer. The final enriched gene set was specifically named 'CD8+ T-cell downregulation', pointing directly to immune system consequences of low BBOX1.
The activation of EMT-related genes in low-BBOX1 tumors is particularly concerning clinically, as EMT is strongly associated with a cancer's ability to invade surrounding tissue and form distant metastases. This helps explain why low BBOX1 tumors tend to present at more advanced stages.
Pathway network analysis further revealed that BBOX1 is linked - indirectly - to several immune-relevant processes, including the T-cell antigen receptor signaling pathway, antigen processing and presentation, and cancer immunotherapy through PD-1 blockade. This positions BBOX1 as a molecule that, when lost, may weaken the immune system's ability to recognize and attack kidney cancer cells.
Analyzing immune cell composition using CIBERSORT, the study found that patients with low BBOX1 expression had a distinctly different immune microenvironment. Most critically, they had significantly fewer CD8+ T cells - the immune cells responsible for directly killing cancer cells (p=0.005). This finding suggests that low BBOX1 tumors are less immunologically active in an anti-cancer sense.
Patients with low BBOX1 also had more neutrophils in their tumors (p=0.048). Tumor-infiltrating neutrophils are generally considered pro-tumorigenic in kidney cancer - they suppress effective immune responses and can promote cancer cell migration and invasion. Higher neutrophil levels are associated with poorer prognosis in RCC.
Low BBOX1 expression was also associated with elevated PD-L1 expression (CD274) and a higher cancer-testis antigen (CTA) score. PD-L1 is one of the key molecules that cancer cells use to hide from the immune system - it is also the target of some immunotherapy drugs. However, the combination of low CD8+ T cells and high PD-L1 in low-BBOX1 tumors may suggest these patients could be resistant to anti-PD-L1 therapy rather than sensitive to it.
Fewer M1 macrophages (the cancer-fighting type) were also found in low BBOX1 tumors (p=0.001). M1 macrophages normally help coordinate immune attacks on cancer cells, so their reduction further depletes the anti-tumor immune capacity in these patients. Together, these immune changes paint a picture of a comprehensively suppressed immune environment in low-BBOX1 kidney cancers.
To identify potential treatments for the subset of kidney cancer patients with low BBOX1 expression, researchers analyzed drug sensitivity data from the Genomics of Drug Sensitivity in Cancer (GDSC) database, covering 30 kidney cancer cell lines. Four drugs showed significantly greater effectiveness against kidney cancer cell lines with low BBOX1 expression.
Midostaurin (also known as PKC412) was the most notable candidate. It is already approved for treating a type of blood cancer with certain genetic mutations and works by inhibiting multiple kinase pathways, blocking tumor blood vessel formation, and triggering cancer cell death. Laboratory evidence in kidney cancer cell lines suggests it targets relevant signaling pathways active in low-BBOX1 tumors.
BAY-61-3606 is a highly selective inhibitor of Syk tyrosine kinase, an enzyme involved in immune signaling and cancer cell survival. It works by causing cell cycle arrest and triggering programmed cell death in cancer cells. GSK690693 targets the AKT/PTEN pathway, which is frequently disrupted in ccRCC and is one of the gene sets enriched in low-BBOX1 tumors - making it a biologically logical candidate.
Linifanib targets VEGF receptors and PDGF receptors - both involved in tumor blood vessel formation (angiogenesis) that is critical for kidney cancer growth. It is similar in mechanism to sunitinib, a standard kidney cancer treatment. The identification of these four drug candidates provides a starting point for developing more targeted treatment strategies for the low-BBOX1 patient subgroup.
BBOX1's normal role in healthy kidney cells is to catalyze the final step in producing L-carnitine - a molecule essential for transporting fatty acids into mitochondria for energy production. Its high expression in normal kidney tubule cells reflects the kidney's high metabolic demands. When BBOX1 is lost in cancer, this metabolic function is disrupted.
Cancer cells frequently reprogram their energy metabolism, shifting away from normal fat oxidation toward glycolysis (the Warburg effect). The loss of BBOX1 may be part of this metabolic reprogramming in ccRCC, contributing to the activation of oncogenic pathways and the altered immune environment observed in low-BBOX1 tumors.
The findings in this study are consistent with reports from the Human Protein Atlas, which also linked low BBOX1 expression to unfavorable prognosis in renal cancer. However, the relationship between BBOX1 and cancer outcomes appears to vary by cancer type - in colorectal cancer, high BBOX1 was linked to higher risk. This highlights how the same gene can play different roles depending on the cancer context.
The connection between BBOX1 and immune function - particularly through pathways linked to T-cell signaling and PD-1/PD-L1 immunotherapy - suggests that BBOX1 may be more than a metabolic enzyme in cancer biology. Understanding how BBOX1 loss shapes the immune environment could lead to insights about why some kidney cancer patients respond poorly to immunotherapy.
This study establishes low BBOX1 expression as a significant independent prognostic biomarker in clear cell kidney cancer. Validated in two separate patient cohorts across different measurement platforms (protein by IHC and RNA by sequencing), the finding that low BBOX1 predicts worse survival is robust and clinically meaningful.
The use of machine learning (gradient boosting machine) demonstrated that BBOX1 adds meaningful information beyond conventional clinical staging and grading systems. This supports the potential clinical utility of incorporating BBOX1 measurement into tumor assessment to improve risk stratification and guide treatment decisions.
The identification of four drug candidates that show enhanced effectiveness against low-BBOX1 kidney cancers opens avenues for developing tailored treatment strategies. With further experimental validation, these drugs - particularly midostaurin - could be advanced into clinical trials specifically targeting low-BBOX1 ccRCC patients.
Limitations include the retrospective nature of the hospital cohort analysis, the absence of functional in vitro or in vivo experiments directly testing how BBOX1 influences immune cell behavior, and the need for prospective validation. Future studies should explore the mechanisms by which BBOX1 loss reshapes the tumor immune microenvironment and test the identified drug candidates in clinical settings.