Investigating Citrate Transporter Utilization and Splicing Patterns Across Solid Tumor Types
Uncovering Tumor Biology Through Genomic Insights
Duration: Fall 2024
Role: Data Scientist Fellow
🚧 🚧
Authors: Tyler Gustafson, Blake Bleier
Research Mentors: Neha Rohatgi, Rohit Reja
Abstract
How does alternative splicing of the pmCiC isoform contribute to its overactive expression and the metabolic reprogramming observed in cancer cells?
Cancer cells exhibit metabolic reprogramming to support their rapid growth and survival. Unlike normal cells that rely primarily on glycolysis and mitochondrial energy production, cancer cells uniquely adapt to utilize extracellular citrate as a key energy and biosynthesis substrate. This adaptation is facilitated by an alternative isoform of the mitochondrial citrate transporter, known as the plasma membrane citrate transporter (pmCiC), which is differentially expressed in cancer cells.
Alternative splicing, a cellular process allowing a single gene to produce multiple mRNA and protein isoforms, plays a pivotal role in generating functional diversity (Process shown in Figure 1). ​
​
While most human genes undergo alternative splicing, cancer cells exploit this process to regulate aberrant isoform expression. This adaptation supports cancer’s rapid growth and division, meeting its high energy demands and providing alternative building materials such as lipids. See Figure 2 for a visualization of this cancer-specific mechanism.
Figure 1: Alternative Splicing
Our study focuses on understanding how differential splicing of pmCiC contributes to its overactive expression, thereby promoting citrate uptake and fueling cancer progression.
Leveraging large-scale cancer genomics data from The Cancer Genome Atlas (TCGA), this project aims to comprehensively investigate key aspects of pmCiC isoform activity, including:
Identify the prevalence of the overactive pmCiC isoform across various cancer types.
Examine differential splicing patterns associated with pmCiC in solid tumors.
Explore the implications of pmCiC activity on cancer cell metabolism and progression.
1.
2.
3.
The Cancer Genome Atlas (TCGA) is a landmark research program that has cataloged genetic mutations responsible for cancer using genome sequencing and bioinformatics. It contains a vast, publicly available dataset covering over 11,000 patients across more than 30 different cancer types, providing an unparalleled resource for studying cancer biology and identifying novel therapeutic targets.​
​
This research provides a foundational understanding of citrate metabolism in cancer cells and establishes a framework for future investigations into therapeutic interventions targeting aberrant splicing and citrate uptake mechanisms.
Research
Concept
Cancer cells exhibit a distinctive metabolic shift to meet their rapid growth and energy demands. Unlike normal cells, which primarily rely on glycolysis to produce pyruvate that enters mitochondria for energy production, cancer cells convert pyruvate to lactate, a hallmark of the Warburg Effect, and depend on alternative energy sources to fuel the tricarboxylic acid (TCA) cycle. One critical source is extracellular citrate, which cancer cells actively acquire from their environment. This process is mediated by an alternative isoform of the mitochondrial citrate transporter, known as the plasma membrane citrate transporter (pmCiC).
As illustrated in the diagram to the right, this adaptation sustains cancer cell metabolism and supports biosynthesis for rapid cell division. By leveraging large-scale cancer genomics data from The Cancer Genome Atlas (TCGA), we aim to investigate how cancer cells exploit this unique metabolic pathway across various cancer types.
Figure 2: Cancer Metabolism: Citrate Mechanism
Approach
Our approach integrates comprehensive data analysis techniques to unravel the role of pmCiC in cancer metabolism. We began with TCGA data ingestion and data preprocessing to ensure clean and structured datasets. This was followed by exploratory data analysis (EDA) to uncover patterns and insights. To validate our hypotheses, we applied statistical testing, complemented by survival analysis and linear deconvolution methods to assess the relationship between pmCiC activity, citrate metabolism, and cancer progression. This multi-step methodology enables a robust investigation of the metabolic reprogramming driven by pmCiC.
Hypothesis Framework
This analysis investigates the role of pmCiC expression in solid tumors by leveraging transcriptomic data from The Cancer Genome Atlas (TCGA). Expression levels were normalized and compared between tumor and normal tissues to identify significant trends. Ratios of pmCiC expression to overall citrate transporter expression were calculated and analyzed across tissue types and tumor stages. Statistical tests evaluated the significance of observed differences, while subgroup and survival analyses explored clinical implications. Below are the three key questions driving our analysis:
-
Q1 | Is there a trend of increased citrate transporter expression in tumors versus normal in TCGA data?
-
Q2 | Is there a significant difference in proportion of pmCiC to mCiC across different tissue types?
-
Q3 | What further insights can we gain, such as subgroup patterns or survival analysis?
Testing for Normality
Before testing, we assessed distribution normality to check for Gaussian (normal) characteristics. The plasma isoform distributions, like others, show a low likelihood of normality. ​
Statistical Approach
To test our hypothesis, we employed a robust statistical framework tailored to non-normal data distributions. This three-step approach examines isoform expression differences across cancerous and normal samples, evaluates variations across cancer types, and investigates specific pairwise comparisons within each cancer type.
Evaluation: Curse of Multiple Testing
Our biggest remaining challenge was to manage the curse of multiple testing. This issue arises from conducting multiple statistical tests, which increases the risk of false positives, and requires careful correction methods to ensure the reliability of our results. In other words, how can we ensure statistical soundness with the challenge of multiple testing due to the ease of generating additional data?
​
To address this, we applied the Benjamini-Hochberg procedure, a widely used method to control the false discovery rate in multiple testing. Unlike the Bonferroni correction, which can be overly conservative and reduce statistical power, the Benjamini-Hochberg approach balances the need to identify true positives while limiting the proportion of false positives. This method was well-suited to the scope of our experiment and hypothesis, ensuring that we maintained statistical rigor without being overly restrictive in detecting meaningful results.
Results
Q1 | Is there a trend of increased citrate transporter expression in tumors versus normal in TCGA data?
General Upregulation in Tumor Tissues
There is statistically significant upregulation of both the mitochondrial citrate transporter (mCiC) and its alternative isoform, the plasma membrane citrate transporter (pmCiC), in tumor samples compared to normal tissues, across a range of tissue types (significance indicated with *). This suggests a general increase in citrate transporter expression associated with cancer, supporting their role in tumor metabolism and growth.
Q2 | Is there a significant difference in proportion of pmCiC to mCiC across different tissue types?
​There are Tissue-Specific Proportional Differences in pmCiC Expression
The proportion of pmCiC to mCiC was statistically significant across two tissue types: lung and breast (all significant findings highlighted in blue to the right). These are fewer tissue types than those showing significant differences in overall expression levels (TpM) above.
​
This suggests that while the citrate transporter is broadly upregulated in cancer, differences in the proportion between pmCiC and mCiC are confined to only a few specific tissue types.
Further exploration revealed that these proportional differences may reflect tissue-specific metabolic demands. Breast and lung tissues, for example, are known for their distinct microenvironments and metabolic phenotypes, which could influence the differential expression of pmCiC relative to mCiC. These findings align with prior research highlighting the metabolic heterogeneity of tumors and underscore the potential for pmCiC to play a more prominent role in certain cancers compared to others.
Q3 | What further insights can we gain, such as subgroup patterns or survival analysis?
Subtype-Specific Expression Patterns in Lung and Breast Cancer
Subgroups were analyzed to isolate the source of differences in pmCiC proportion. Many subgroup types were considered, including tumor stage, disease-specific survival, mRNA subtype, and immune subtype, among others. In breast cancer, elevated statistically significant differences were observed in Basal and Normal mRNA subtypes and the Inflammatory immune subtype, while in lung cancer, significance was observed solely in the C1 (Wound Healing) immune subtype.
Unexpected Survival Trends Suggest Limited Prognostic Impact
Survival analysis was performed on both subtype-specific analyses and overall pmCiC proportion groups. Contrary to the initial hypothesis, subgroups with higher pmCiC proportions showed a qualitative trend toward longer survival. Statistically, there were limited and non-significant differences in survival among these subgroups, suggesting that elevated pmCiC proportions may not be associated with decreased survival as expected.
Tumor-Stromal Interplay Unclear in Linear Deconvolution
A linear deconvolution analysis was conducted to determine the contribution of tumor and stromal cells to the expression of mCiC and pmCiC across varying tumor purities. No observable trend was found for mCiC or pmCiC in breast tissue. In lung tissue, a general upregulation of citrate transporters was observed, but without evidence of specific upregulation of pmCiC, suggesting that its relative proportion may not be a key factor in this tissue.
Discussion & Broader Applications
Understanding the Role of Splicing in Tumor Metabolism
Our study sheds light on the tissue-specific metabolic reprogramming driven by alternative splicing of citrate transporters, with a focus on pmCiC. By combining statistical analyses with insights from survival and deconvolution studies, we provide a nuanced perspective on how pmCiC contributes to tumor progression. Our findings enhance the understanding of cancer metabolism and present new opportunities for therapeutic interventions targeting aberrant splicing in specific cancer contexts.
The methodologies and insights from this project have broader potential applications in other cancer research and beyond:
-
Generalizable Downstream Pipeline: The approach can be adapted to study alternate splice variant upregulation for other genes of interest.
-
Systematic Framework: Provides a scalable framework to identify and analyze alternate splicing events in large-scale cancer datasets.
-
Cross-Cancer Comparisons: Enable comparisons across cancer types to identify shared or unique splicing patterns
Future Research
​Building on the findings of this study, future research can focus on validating and expanding the current work to gain deeper insights into the role of pmCiC in cancer metabolism. These next steps aim to strengthen our understanding of citrate transporter function and explore their potential as therapeutic targets in specific cancer types.
-
Long-Read Sequencing in Breast/Lung Cancer: Investigate mitochondrial or plasma membrane transcript presence using long-read sequencing to better understand their relative abundances and validate our hypothesis.
-
Explore Additional Datasets: Utilize independent datasets like ICGC or cancer cell line datasets (e.g., CCLE) to further test and strengthen the hypothesis.
-
Functional Validation: Design experiments to assess the functional relevance of these splice variants in cancer models.
Acknowledgements
The results shown here are in whole or part based upon data generated by the TCGA Research network: https://www.cancer.gov/tcga
Additionally, this research builds upon findings from the study:
"Pan-Cancer Analysis of Ligand–Receptor Cross-talk in the Tumor Microenvironment" (Cancer Res, 2021, 81(7):1802–1812) by Ghoshdastider et al.
A special thank you to Rohit Reja and Neha Rohatgi for their guidance and expertise as scientific advisors throughout this project. Their insights on experimental design and future directions have been instrumental in shaping our work.
